中級者向け No.079
機械学習デモアプリ
CSVデータをscikit-learnで分類・回帰分析し結果と精度をグラフで可視化するML入門デモアプリ。
📋 目次
1. アプリ概要
CSVデータをscikit-learnで分類・回帰分析し結果と精度をグラフで可視化するML入門デモアプリ。
このアプリはdataカテゴリの実践的なPythonアプリです。使用ライブラリは tkinter（標準ライブラリ）・pandas・matplotlib、難易度は ★★★ です。
Pythonの豊富なライブラリを活用することで、実用的なアプリを短いコードで実装できます。ソースコードをコピーして実行し、仕組みを理解したうえでカスタマイズに挑戦してみてください。
GUIアプリ開発はプログラミングの楽しさを実感できる最も効果的な学習方法のひとつです。変数・関数・クラス・イベント処理などの重要な概念が自然と身につきます。
2. 機能一覧
- 機械学習デモアプリのメイン機能
- 直感的なGUIインターフェース
- 入力値のバリデーション
- エラーハンドリング
- 結果の見やすい表示
- クリア機能付き
3. 事前準備・環境
動作確認環境
Python 3.10 以上 / Windows・Mac・Linux すべて対応
以下の環境で動作確認しています。
- Python 3.10 以上
- OS: Windows 10/11・macOS 12+・Ubuntu 20.04+
インストールが必要なライブラリ
pip install scikit-learn matplotlib numpy4. 完全なソースコード
コードのコピー方法
右上の「コピー」ボタンをクリックするとコードをクリップボードにコピーできます。
app079.py
import tkinter as tk
from tkinter import ttk, messagebox
import threading
import random
import math
try:
import matplotlib
matplotlib.use("TkAgg")
from matplotlib.figure import Figure
from matplotlib.backends.backend_tkagg import FigureCanvasTkAgg
MATPLOTLIB_AVAILABLE = True
except ImportError:
MATPLOTLIB_AVAILABLE = False
try:
from sklearn.datasets import make_classification, make_regression, make_blobs
from sklearn.model_selection import train_test_split
from sklearn.preprocessing import StandardScaler
from sklearn.linear_model import LogisticRegression, LinearRegression
from sklearn.tree import DecisionTreeClassifier
from sklearn.ensemble import RandomForestClassifier
from sklearn.svm import SVC
from sklearn.neighbors import KNeighborsClassifier
from sklearn.metrics import (accuracy_score, classification_report,
mean_squared_error, r2_score)
import numpy as np
SKLEARN_AVAILABLE = True
except ImportError:
SKLEARN_AVAILABLE = False
class App079:
"""機械学習デモアプリ"""
CLASSIFIERS = {
"ロジスティック回帰": "LogisticRegression",
"決定木": "DecisionTreeClassifier",
"ランダムフォレスト": "RandomForestClassifier",
"SVM": "SVC",
"k近傍法": "KNeighborsClassifier",
}
def __init__(self, root):
self.root = root
self.root.title("機械学習デモアプリ")
self.root.geometry("1000x640")
self.root.configure(bg="#1e1e1e")
self._build_ui()
def _build_ui(self):
header = tk.Frame(self.root, bg="#252526", pady=6)
header.pack(fill=tk.X)
tk.Label(header, text="🤖 機械学習デモアプリ",
font=("Noto Sans JP", 12, "bold"),
bg="#252526", fg="#4fc3f7").pack(side=tk.LEFT, padx=12)
missing = []
if not SKLEARN_AVAILABLE: missing.append("scikit-learn")
if not MATPLOTLIB_AVAILABLE: missing.append("matplotlib")
if missing:
tk.Label(self.root,
text=f"⚠ pip install {' '.join(missing)}",
bg="#fff3cd", fg="#856404", font=("Arial", 9),
anchor="w", padx=8).pack(fill=tk.X)
# 設定パネル
cfg = tk.LabelFrame(self.root, text="設定", bg="#252526",
fg="#ccc", font=("Arial", 9), padx=8, pady=6)
cfg.pack(fill=tk.X, padx=8, pady=4)
# タスク
tk.Label(cfg, text="タスク:", bg="#252526", fg="#ccc",
font=("Arial", 9)).grid(row=0, column=0, sticky="w")
self.task_var = tk.StringVar(value="分類")
ttk.Combobox(cfg, textvariable=self.task_var,
values=["分類", "回帰", "クラスタリング"],
state="readonly", width=12).grid(row=0, column=1, padx=4)
# アルゴリズム
tk.Label(cfg, text="アルゴリズム:", bg="#252526", fg="#ccc",
font=("Arial", 9)).grid(row=0, column=2, sticky="w", padx=(12, 0))
self.algo_var = tk.StringVar(value="ロジスティック回帰")
self.algo_cb = ttk.Combobox(cfg, textvariable=self.algo_var,
values=list(self.CLASSIFIERS.keys()),
state="readonly", width=18)
self.algo_cb.grid(row=0, column=3, padx=4)
# サンプル数
tk.Label(cfg, text="サンプル数:", bg="#252526", fg="#ccc",
font=("Arial", 9)).grid(row=0, column=4, sticky="w", padx=(12, 0))
self.n_samples_var = tk.IntVar(value=300)
ttk.Spinbox(cfg, from_=50, to=2000, textvariable=self.n_samples_var,
width=6).grid(row=0, column=5, padx=4)
# テスト割合
tk.Label(cfg, text="テスト割合:", bg="#252526", fg="#ccc",
font=("Arial", 9)).grid(row=0, column=6, sticky="w", padx=(8, 0))
self.test_size_var = tk.DoubleVar(value=0.2)
ttk.Spinbox(cfg, from_=0.1, to=0.5, increment=0.05,
textvariable=self.test_size_var,
width=5).grid(row=0, column=7, padx=4)
tk.Button(cfg, text="▶ 実行", command=self._run,
bg="#1565c0", fg="white", relief=tk.FLAT,
font=("Arial", 10, "bold"), padx=14, pady=4,
activebackground="#0d47a1", bd=0).grid(
row=0, column=8, padx=8)
self.task_var.trace_add("write", self._on_task_change)
# メイン: 左=グラフ, 右=レポート
main = tk.Frame(self.root, bg="#1e1e1e")
main.pack(fill=tk.BOTH, expand=True, padx=8, pady=4)
# 左: グラフ
left = tk.Frame(main, bg="#0d1117")
left.pack(side=tk.LEFT, fill=tk.BOTH, expand=True)
if MATPLOTLIB_AVAILABLE:
self.fig = Figure(figsize=(6, 5), facecolor="#0d1117")
self.ax = self.fig.add_subplot(111)
self.mpl_canvas = FigureCanvasTkAgg(self.fig, master=left)
self.mpl_canvas.get_tk_widget().pack(fill=tk.BOTH, expand=True)
# 右: 結果レポート
right = tk.Frame(main, bg="#1e1e1e", width=340)
right.pack(side=tk.LEFT, fill=tk.Y, padx=(8, 0))
right.pack_propagate(False)
tk.Label(right, text="評価レポート", bg="#1e1e1e", fg="#888",
font=("Arial", 9)).pack(anchor="w")
self.report_text = tk.Text(right, bg="#0d1117", fg="#c9d1d9",
font=("Courier New", 9), relief=tk.FLAT,
state=tk.DISABLED)
self.report_text.pack(fill=tk.BOTH, expand=True)
self.status_var = tk.StringVar(value="「実行」ボタンで機械学習デモを開始します")
tk.Label(self.root, textvariable=self.status_var,
bg="#252526", fg="#858585", font=("Arial", 9),
anchor="w", padx=8).pack(fill=tk.X, side=tk.BOTTOM)
def _on_task_change(self, *_):
task = self.task_var.get()
if task == "分類":
self.algo_cb["values"] = list(self.CLASSIFIERS.keys())
self.algo_var.set("ロジスティック回帰")
elif task == "回帰":
self.algo_cb["values"] = ["線形回帰"]
self.algo_var.set("線形回帰")
else:
self.algo_cb["values"] = ["K-Means"]
self.algo_var.set("K-Means")
def _run(self):
if not SKLEARN_AVAILABLE or not MATPLOTLIB_AVAILABLE:
messagebox.showerror("エラー",
"pip install scikit-learn matplotlib")
return
self.status_var.set("実行中...")
threading.Thread(target=self._do_run, daemon=True).start()
def _do_run(self):
try:
task = self.task_var.get()
n = self.n_samples_var.get()
ts = self.test_size_var.get()
algo = self.algo_var.get()
if task == "分類":
self._run_classification(n, ts, algo)
elif task == "回帰":
self._run_regression(n, ts)
else:
self._run_clustering(n)
except Exception as e:
self.root.after(0, self.status_var.set, f"エラー: {e}")
def _run_classification(self, n, ts, algo_name):
X, y = make_classification(
n_samples=n, n_features=2, n_informative=2,
n_redundant=0, n_clusters_per_class=1, random_state=42)
scaler = StandardScaler()
X_s = scaler.fit_transform(X)
X_tr, X_te, y_tr, y_te = train_test_split(X_s, y, test_size=ts, random_state=42)
models = {
"ロジスティック回帰": LogisticRegression(max_iter=200),
"決定木": DecisionTreeClassifier(max_depth=5),
"ランダムフォレスト": RandomForestClassifier(n_estimators=100),
"SVM": SVC(kernel="rbf", probability=True),
"k近傍法": KNeighborsClassifier(n_neighbors=5),
}
model = models.get(algo_name, LogisticRegression())
model.fit(X_tr, y_tr)
y_pred = model.predict(X_te)
acc = accuracy_score(y_te, y_pred)
report = classification_report(y_te, y_pred, target_names=["クラス0", "クラス1"])
self.root.after(0, self._render_classification,
X_s, y, model, acc, report, algo_name)
def _render_classification(self, X, y, model, acc, report, algo_name):
self.ax.clear()
self.ax.set_facecolor("#0d1117")
# 決定境界
h = 0.05
x_min, x_max = X[:, 0].min() - 0.5, X[:, 0].max() + 0.5
y_min, y_max = X[:, 1].min() - 0.5, X[:, 1].max() + 0.5
xx, yy = np.meshgrid(np.arange(x_min, x_max, h),
np.arange(y_min, y_max, h))
Z = model.predict(np.c_[xx.ravel(), yy.ravel()])
Z = Z.reshape(xx.shape)
self.ax.contourf(xx, yy, Z, alpha=0.25,
cmap="RdYlGn", levels=[-0.5, 0.5, 1.5])
colors = ["#ef5350", "#26a69a"]
for cls in [0, 1]:
mask = y == cls
self.ax.scatter(X[mask, 0], X[mask, 1], color=colors[cls],
s=18, alpha=0.7, label=f"クラス{cls}")
self.ax.set_title(f"{algo_name} 精度: {acc:.3f}", color="#c9d1d9")
self.ax.tick_params(colors="#8b949e", labelsize=7)
for spine in self.ax.spines.values():
spine.set_edgecolor("#30363d")
self.ax.legend(facecolor="#161b22", labelcolor="#c9d1d9",
edgecolor="#30363d", fontsize=8)
self.fig.tight_layout()
self.mpl_canvas.draw()
self._set_report(f"アルゴリズム: {algo_name}\n精度: {acc:.4f}\n\n{report}")
self.status_var.set(f"完了: {algo_name} 精度 {acc:.4f}")
def _run_regression(self, n, ts):
X, y = make_regression(n_samples=n, n_features=1, noise=20, random_state=42)
X_tr, X_te, y_tr, y_te = train_test_split(X, y, test_size=ts, random_state=42)
model = LinearRegression()
model.fit(X_tr, y_tr)
y_pred = model.predict(X_te)
mse = mean_squared_error(y_te, y_pred)
r2 = r2_score(y_te, y_pred)
self.root.after(0, self._render_regression,
X, y, model, mse, r2)
def _render_regression(self, X, y, model, mse, r2):
self.ax.clear()
self.ax.set_facecolor("#0d1117")
self.ax.scatter(X, y, color="#4fc3f7", s=15, alpha=0.6, label="データ")
x_line = np.linspace(X.min(), X.max(), 100).reshape(-1, 1)
y_line = model.predict(x_line)
self.ax.plot(x_line, y_line, color="#ef5350", linewidth=2, label="予測")
self.ax.set_title(f"線形回帰 R²: {r2:.3f}", color="#c9d1d9")
self.ax.tick_params(colors="#8b949e", labelsize=7)
for spine in self.ax.spines.values():
spine.set_edgecolor("#30363d")
self.ax.legend(facecolor="#161b22", labelcolor="#c9d1d9",
edgecolor="#30363d", fontsize=8)
self.fig.tight_layout()
self.mpl_canvas.draw()
self._set_report(
f"アルゴリズム: 線形回帰\n"
f"MSE: {mse:.4f}\nRMSE: {math.sqrt(mse):.4f}\nR²: {r2:.4f}\n"
f"係数: {model.coef_[0]:.4f}\n切片: {model.intercept_:.4f}")
self.status_var.set(f"完了: 線形回帰 R² {r2:.4f}")
def _run_clustering(self, n):
from sklearn.cluster import KMeans
X, _ = make_blobs(n_samples=n, centers=4, random_state=42)
kmeans = KMeans(n_clusters=4, random_state=42, n_init=10)
labels = kmeans.fit_predict(X)
self.root.after(0, self._render_clustering, X, labels, kmeans)
def _render_clustering(self, X, labels, kmeans):
self.ax.clear()
self.ax.set_facecolor("#0d1117")
colors = ["#4fc3f7", "#ef5350", "#26a69a", "#ffa726"]
for i in range(4):
mask = labels == i
self.ax.scatter(X[mask, 0], X[mask, 1], color=colors[i],
s=15, alpha=0.6, label=f"クラスタ{i}")
centers = kmeans.cluster_centers_
self.ax.scatter(centers[:, 0], centers[:, 1],
marker="*", s=200, color="white", zorder=5)
self.ax.set_title("K-Means クラスタリング (k=4)", color="#c9d1d9")
self.ax.tick_params(colors="#8b949e", labelsize=7)
for spine in self.ax.spines.values():
spine.set_edgecolor("#30363d")
self.ax.legend(facecolor="#161b22", labelcolor="#c9d1d9",
edgecolor="#30363d", fontsize=8)
self.fig.tight_layout()
self.mpl_canvas.draw()
inertia = kmeans.inertia_
self._set_report(
f"アルゴリズム: K-Means\nクラスタ数: 4\n"
f"慣性 (inertia): {inertia:.2f}\nサンプル数: {len(X)}")
self.status_var.set(f"完了: K-Means 慣性 {inertia:.2f}")
def _set_report(self, text):
self.report_text.configure(state=tk.NORMAL)
self.report_text.delete("1.0", tk.END)
self.report_text.insert("1.0", text)
self.report_text.configure(state=tk.DISABLED)
if __name__ == "__main__":
root = tk.Tk()
app = App079(root)
root.mainloop()
5. コード解説
機械学習デモアプリのコードを詳しく解説します。クラスベースの設計で各機能を整理して実装しています。
クラス設計とコンストラクタ
App079クラスにアプリの全機能をまとめています。__init__でウィンドウ設定、_build_ui()でUI構築、process()でメイン処理を担当します。責任の分離により、コードが読みやすくなります。
import tkinter as tk
from tkinter import ttk, messagebox
import threading
import random
import math
try:
import matplotlib
matplotlib.use("TkAgg")
from matplotlib.figure import Figure
from matplotlib.backends.backend_tkagg import FigureCanvasTkAgg
MATPLOTLIB_AVAILABLE = True
except ImportError:
MATPLOTLIB_AVAILABLE = False
try:
from sklearn.datasets import make_classification, make_regression, make_blobs
from sklearn.model_selection import train_test_split
from sklearn.preprocessing import StandardScaler
from sklearn.linear_model import LogisticRegression, LinearRegression
from sklearn.tree import DecisionTreeClassifier
from sklearn.ensemble import RandomForestClassifier
from sklearn.svm import SVC
from sklearn.neighbors import KNeighborsClassifier
from sklearn.metrics import (accuracy_score, classification_report,
mean_squared_error, r2_score)
import numpy as np
SKLEARN_AVAILABLE = True
except ImportError:
SKLEARN_AVAILABLE = False
class App079:
"""機械学習デモアプリ"""
CLASSIFIERS = {
"ロジスティック回帰": "LogisticRegression",
"決定木": "DecisionTreeClassifier",
"ランダムフォレスト": "RandomForestClassifier",
"SVM": "SVC",
"k近傍法": "KNeighborsClassifier",
}
def __init__(self, root):
self.root = root
self.root.title("機械学習デモアプリ")
self.root.geometry("1000x640")
self.root.configure(bg="#1e1e1e")
self._build_ui()
def _build_ui(self):
header = tk.Frame(self.root, bg="#252526", pady=6)
header.pack(fill=tk.X)
tk.Label(header, text="🤖 機械学習デモアプリ",
font=("Noto Sans JP", 12, "bold"),
bg="#252526", fg="#4fc3f7").pack(side=tk.LEFT, padx=12)
missing = []
if not SKLEARN_AVAILABLE: missing.append("scikit-learn")
if not MATPLOTLIB_AVAILABLE: missing.append("matplotlib")
if missing:
tk.Label(self.root,
text=f"⚠ pip install {' '.join(missing)}",
bg="#fff3cd", fg="#856404", font=("Arial", 9),
anchor="w", padx=8).pack(fill=tk.X)
# 設定パネル
cfg = tk.LabelFrame(self.root, text="設定", bg="#252526",
fg="#ccc", font=("Arial", 9), padx=8, pady=6)
cfg.pack(fill=tk.X, padx=8, pady=4)
# タスク
tk.Label(cfg, text="タスク:", bg="#252526", fg="#ccc",
font=("Arial", 9)).grid(row=0, column=0, sticky="w")
self.task_var = tk.StringVar(value="分類")
ttk.Combobox(cfg, textvariable=self.task_var,
values=["分類", "回帰", "クラスタリング"],
state="readonly", width=12).grid(row=0, column=1, padx=4)
# アルゴリズム
tk.Label(cfg, text="アルゴリズム:", bg="#252526", fg="#ccc",
font=("Arial", 9)).grid(row=0, column=2, sticky="w", padx=(12, 0))
self.algo_var = tk.StringVar(value="ロジスティック回帰")
self.algo_cb = ttk.Combobox(cfg, textvariable=self.algo_var,
values=list(self.CLASSIFIERS.keys()),
state="readonly", width=18)
self.algo_cb.grid(row=0, column=3, padx=4)
# サンプル数
tk.Label(cfg, text="サンプル数:", bg="#252526", fg="#ccc",
font=("Arial", 9)).grid(row=0, column=4, sticky="w", padx=(12, 0))
self.n_samples_var = tk.IntVar(value=300)
ttk.Spinbox(cfg, from_=50, to=2000, textvariable=self.n_samples_var,
width=6).grid(row=0, column=5, padx=4)
# テスト割合
tk.Label(cfg, text="テスト割合:", bg="#252526", fg="#ccc",
font=("Arial", 9)).grid(row=0, column=6, sticky="w", padx=(8, 0))
self.test_size_var = tk.DoubleVar(value=0.2)
ttk.Spinbox(cfg, from_=0.1, to=0.5, increment=0.05,
textvariable=self.test_size_var,
width=5).grid(row=0, column=7, padx=4)
tk.Button(cfg, text="▶ 実行", command=self._run,
bg="#1565c0", fg="white", relief=tk.FLAT,
font=("Arial", 10, "bold"), padx=14, pady=4,
activebackground="#0d47a1", bd=0).grid(
row=0, column=8, padx=8)
self.task_var.trace_add("write", self._on_task_change)
# メイン: 左=グラフ, 右=レポート
main = tk.Frame(self.root, bg="#1e1e1e")
main.pack(fill=tk.BOTH, expand=True, padx=8, pady=4)
# 左: グラフ
left = tk.Frame(main, bg="#0d1117")
left.pack(side=tk.LEFT, fill=tk.BOTH, expand=True)
if MATPLOTLIB_AVAILABLE:
self.fig = Figure(figsize=(6, 5), facecolor="#0d1117")
self.ax = self.fig.add_subplot(111)
self.mpl_canvas = FigureCanvasTkAgg(self.fig, master=left)
self.mpl_canvas.get_tk_widget().pack(fill=tk.BOTH, expand=True)
# 右: 結果レポート
right = tk.Frame(main, bg="#1e1e1e", width=340)
right.pack(side=tk.LEFT, fill=tk.Y, padx=(8, 0))
right.pack_propagate(False)
tk.Label(right, text="評価レポート", bg="#1e1e1e", fg="#888",
font=("Arial", 9)).pack(anchor="w")
self.report_text = tk.Text(right, bg="#0d1117", fg="#c9d1d9",
font=("Courier New", 9), relief=tk.FLAT,
state=tk.DISABLED)
self.report_text.pack(fill=tk.BOTH, expand=True)
self.status_var = tk.StringVar(value="「実行」ボタンで機械学習デモを開始します")
tk.Label(self.root, textvariable=self.status_var,
bg="#252526", fg="#858585", font=("Arial", 9),
anchor="w", padx=8).pack(fill=tk.X, side=tk.BOTTOM)
def _on_task_change(self, *_):
task = self.task_var.get()
if task == "分類":
self.algo_cb["values"] = list(self.CLASSIFIERS.keys())
self.algo_var.set("ロジスティック回帰")
elif task == "回帰":
self.algo_cb["values"] = ["線形回帰"]
self.algo_var.set("線形回帰")
else:
self.algo_cb["values"] = ["K-Means"]
self.algo_var.set("K-Means")
def _run(self):
if not SKLEARN_AVAILABLE or not MATPLOTLIB_AVAILABLE:
messagebox.showerror("エラー",
"pip install scikit-learn matplotlib")
return
self.status_var.set("実行中...")
threading.Thread(target=self._do_run, daemon=True).start()
def _do_run(self):
try:
task = self.task_var.get()
n = self.n_samples_var.get()
ts = self.test_size_var.get()
algo = self.algo_var.get()
if task == "分類":
self._run_classification(n, ts, algo)
elif task == "回帰":
self._run_regression(n, ts)
else:
self._run_clustering(n)
except Exception as e:
self.root.after(0, self.status_var.set, f"エラー: {e}")
def _run_classification(self, n, ts, algo_name):
X, y = make_classification(
n_samples=n, n_features=2, n_informative=2,
n_redundant=0, n_clusters_per_class=1, random_state=42)
scaler = StandardScaler()
X_s = scaler.fit_transform(X)
X_tr, X_te, y_tr, y_te = train_test_split(X_s, y, test_size=ts, random_state=42)
models = {
"ロジスティック回帰": LogisticRegression(max_iter=200),
"決定木": DecisionTreeClassifier(max_depth=5),
"ランダムフォレスト": RandomForestClassifier(n_estimators=100),
"SVM": SVC(kernel="rbf", probability=True),
"k近傍法": KNeighborsClassifier(n_neighbors=5),
}
model = models.get(algo_name, LogisticRegression())
model.fit(X_tr, y_tr)
y_pred = model.predict(X_te)
acc = accuracy_score(y_te, y_pred)
report = classification_report(y_te, y_pred, target_names=["クラス0", "クラス1"])
self.root.after(0, self._render_classification,
X_s, y, model, acc, report, algo_name)
def _render_classification(self, X, y, model, acc, report, algo_name):
self.ax.clear()
self.ax.set_facecolor("#0d1117")
# 決定境界
h = 0.05
x_min, x_max = X[:, 0].min() - 0.5, X[:, 0].max() + 0.5
y_min, y_max = X[:, 1].min() - 0.5, X[:, 1].max() + 0.5
xx, yy = np.meshgrid(np.arange(x_min, x_max, h),
np.arange(y_min, y_max, h))
Z = model.predict(np.c_[xx.ravel(), yy.ravel()])
Z = Z.reshape(xx.shape)
self.ax.contourf(xx, yy, Z, alpha=0.25,
cmap="RdYlGn", levels=[-0.5, 0.5, 1.5])
colors = ["#ef5350", "#26a69a"]
for cls in [0, 1]:
mask = y == cls
self.ax.scatter(X[mask, 0], X[mask, 1], color=colors[cls],
s=18, alpha=0.7, label=f"クラス{cls}")
self.ax.set_title(f"{algo_name} 精度: {acc:.3f}", color="#c9d1d9")
self.ax.tick_params(colors="#8b949e", labelsize=7)
for spine in self.ax.spines.values():
spine.set_edgecolor("#30363d")
self.ax.legend(facecolor="#161b22", labelcolor="#c9d1d9",
edgecolor="#30363d", fontsize=8)
self.fig.tight_layout()
self.mpl_canvas.draw()
self._set_report(f"アルゴリズム: {algo_name}\n精度: {acc:.4f}\n\n{report}")
self.status_var.set(f"完了: {algo_name} 精度 {acc:.4f}")
def _run_regression(self, n, ts):
X, y = make_regression(n_samples=n, n_features=1, noise=20, random_state=42)
X_tr, X_te, y_tr, y_te = train_test_split(X, y, test_size=ts, random_state=42)
model = LinearRegression()
model.fit(X_tr, y_tr)
y_pred = model.predict(X_te)
mse = mean_squared_error(y_te, y_pred)
r2 = r2_score(y_te, y_pred)
self.root.after(0, self._render_regression,
X, y, model, mse, r2)
def _render_regression(self, X, y, model, mse, r2):
self.ax.clear()
self.ax.set_facecolor("#0d1117")
self.ax.scatter(X, y, color="#4fc3f7", s=15, alpha=0.6, label="データ")
x_line = np.linspace(X.min(), X.max(), 100).reshape(-1, 1)
y_line = model.predict(x_line)
self.ax.plot(x_line, y_line, color="#ef5350", linewidth=2, label="予測")
self.ax.set_title(f"線形回帰 R²: {r2:.3f}", color="#c9d1d9")
self.ax.tick_params(colors="#8b949e", labelsize=7)
for spine in self.ax.spines.values():
spine.set_edgecolor("#30363d")
self.ax.legend(facecolor="#161b22", labelcolor="#c9d1d9",
edgecolor="#30363d", fontsize=8)
self.fig.tight_layout()
self.mpl_canvas.draw()
self._set_report(
f"アルゴリズム: 線形回帰\n"
f"MSE: {mse:.4f}\nRMSE: {math.sqrt(mse):.4f}\nR²: {r2:.4f}\n"
f"係数: {model.coef_[0]:.4f}\n切片: {model.intercept_:.4f}")
self.status_var.set(f"完了: 線形回帰 R² {r2:.4f}")
def _run_clustering(self, n):
from sklearn.cluster import KMeans
X, _ = make_blobs(n_samples=n, centers=4, random_state=42)
kmeans = KMeans(n_clusters=4, random_state=42, n_init=10)
labels = kmeans.fit_predict(X)
self.root.after(0, self._render_clustering, X, labels, kmeans)
def _render_clustering(self, X, labels, kmeans):
self.ax.clear()
self.ax.set_facecolor("#0d1117")
colors = ["#4fc3f7", "#ef5350", "#26a69a", "#ffa726"]
for i in range(4):
mask = labels == i
self.ax.scatter(X[mask, 0], X[mask, 1], color=colors[i],
s=15, alpha=0.6, label=f"クラスタ{i}")
centers = kmeans.cluster_centers_
self.ax.scatter(centers[:, 0], centers[:, 1],
marker="*", s=200, color="white", zorder=5)
self.ax.set_title("K-Means クラスタリング (k=4)", color="#c9d1d9")
self.ax.tick_params(colors="#8b949e", labelsize=7)
for spine in self.ax.spines.values():
spine.set_edgecolor("#30363d")
self.ax.legend(facecolor="#161b22", labelcolor="#c9d1d9",
edgecolor="#30363d", fontsize=8)
self.fig.tight_layout()
self.mpl_canvas.draw()
inertia = kmeans.inertia_
self._set_report(
f"アルゴリズム: K-Means\nクラスタ数: 4\n"
f"慣性 (inertia): {inertia:.2f}\nサンプル数: {len(X)}")
self.status_var.set(f"完了: K-Means 慣性 {inertia:.2f}")
def _set_report(self, text):
self.report_text.configure(state=tk.NORMAL)
self.report_text.delete("1.0", tk.END)
self.report_text.insert("1.0", text)
self.report_text.configure(state=tk.DISABLED)
if __name__ == "__main__":
root = tk.Tk()
app = App079(root)
root.mainloop()
UIレイアウトの構築
LabelFrameで入力エリアと結果エリアを視覚的に分けています。pack()で縦に並べ、expand=Trueで結果エリアが画面いっぱいに広がるよう設定しています。
import tkinter as tk
from tkinter import ttk, messagebox
import threading
import random
import math
try:
import matplotlib
matplotlib.use("TkAgg")
from matplotlib.figure import Figure
from matplotlib.backends.backend_tkagg import FigureCanvasTkAgg
MATPLOTLIB_AVAILABLE = True
except ImportError:
MATPLOTLIB_AVAILABLE = False
try:
from sklearn.datasets import make_classification, make_regression, make_blobs
from sklearn.model_selection import train_test_split
from sklearn.preprocessing import StandardScaler
from sklearn.linear_model import LogisticRegression, LinearRegression
from sklearn.tree import DecisionTreeClassifier
from sklearn.ensemble import RandomForestClassifier
from sklearn.svm import SVC
from sklearn.neighbors import KNeighborsClassifier
from sklearn.metrics import (accuracy_score, classification_report,
mean_squared_error, r2_score)
import numpy as np
SKLEARN_AVAILABLE = True
except ImportError:
SKLEARN_AVAILABLE = False
class App079:
"""機械学習デモアプリ"""
CLASSIFIERS = {
"ロジスティック回帰": "LogisticRegression",
"決定木": "DecisionTreeClassifier",
"ランダムフォレスト": "RandomForestClassifier",
"SVM": "SVC",
"k近傍法": "KNeighborsClassifier",
}
def __init__(self, root):
self.root = root
self.root.title("機械学習デモアプリ")
self.root.geometry("1000x640")
self.root.configure(bg="#1e1e1e")
self._build_ui()
def _build_ui(self):
header = tk.Frame(self.root, bg="#252526", pady=6)
header.pack(fill=tk.X)
tk.Label(header, text="🤖 機械学習デモアプリ",
font=("Noto Sans JP", 12, "bold"),
bg="#252526", fg="#4fc3f7").pack(side=tk.LEFT, padx=12)
missing = []
if not SKLEARN_AVAILABLE: missing.append("scikit-learn")
if not MATPLOTLIB_AVAILABLE: missing.append("matplotlib")
if missing:
tk.Label(self.root,
text=f"⚠ pip install {' '.join(missing)}",
bg="#fff3cd", fg="#856404", font=("Arial", 9),
anchor="w", padx=8).pack(fill=tk.X)
# 設定パネル
cfg = tk.LabelFrame(self.root, text="設定", bg="#252526",
fg="#ccc", font=("Arial", 9), padx=8, pady=6)
cfg.pack(fill=tk.X, padx=8, pady=4)
# タスク
tk.Label(cfg, text="タスク:", bg="#252526", fg="#ccc",
font=("Arial", 9)).grid(row=0, column=0, sticky="w")
self.task_var = tk.StringVar(value="分類")
ttk.Combobox(cfg, textvariable=self.task_var,
values=["分類", "回帰", "クラスタリング"],
state="readonly", width=12).grid(row=0, column=1, padx=4)
# アルゴリズム
tk.Label(cfg, text="アルゴリズム:", bg="#252526", fg="#ccc",
font=("Arial", 9)).grid(row=0, column=2, sticky="w", padx=(12, 0))
self.algo_var = tk.StringVar(value="ロジスティック回帰")
self.algo_cb = ttk.Combobox(cfg, textvariable=self.algo_var,
values=list(self.CLASSIFIERS.keys()),
state="readonly", width=18)
self.algo_cb.grid(row=0, column=3, padx=4)
# サンプル数
tk.Label(cfg, text="サンプル数:", bg="#252526", fg="#ccc",
font=("Arial", 9)).grid(row=0, column=4, sticky="w", padx=(12, 0))
self.n_samples_var = tk.IntVar(value=300)
ttk.Spinbox(cfg, from_=50, to=2000, textvariable=self.n_samples_var,
width=6).grid(row=0, column=5, padx=4)
# テスト割合
tk.Label(cfg, text="テスト割合:", bg="#252526", fg="#ccc",
font=("Arial", 9)).grid(row=0, column=6, sticky="w", padx=(8, 0))
self.test_size_var = tk.DoubleVar(value=0.2)
ttk.Spinbox(cfg, from_=0.1, to=0.5, increment=0.05,
textvariable=self.test_size_var,
width=5).grid(row=0, column=7, padx=4)
tk.Button(cfg, text="▶ 実行", command=self._run,
bg="#1565c0", fg="white", relief=tk.FLAT,
font=("Arial", 10, "bold"), padx=14, pady=4,
activebackground="#0d47a1", bd=0).grid(
row=0, column=8, padx=8)
self.task_var.trace_add("write", self._on_task_change)
# メイン: 左=グラフ, 右=レポート
main = tk.Frame(self.root, bg="#1e1e1e")
main.pack(fill=tk.BOTH, expand=True, padx=8, pady=4)
# 左: グラフ
left = tk.Frame(main, bg="#0d1117")
left.pack(side=tk.LEFT, fill=tk.BOTH, expand=True)
if MATPLOTLIB_AVAILABLE:
self.fig = Figure(figsize=(6, 5), facecolor="#0d1117")
self.ax = self.fig.add_subplot(111)
self.mpl_canvas = FigureCanvasTkAgg(self.fig, master=left)
self.mpl_canvas.get_tk_widget().pack(fill=tk.BOTH, expand=True)
# 右: 結果レポート
right = tk.Frame(main, bg="#1e1e1e", width=340)
right.pack(side=tk.LEFT, fill=tk.Y, padx=(8, 0))
right.pack_propagate(False)
tk.Label(right, text="評価レポート", bg="#1e1e1e", fg="#888",
font=("Arial", 9)).pack(anchor="w")
self.report_text = tk.Text(right, bg="#0d1117", fg="#c9d1d9",
font=("Courier New", 9), relief=tk.FLAT,
state=tk.DISABLED)
self.report_text.pack(fill=tk.BOTH, expand=True)
self.status_var = tk.StringVar(value="「実行」ボタンで機械学習デモを開始します")
tk.Label(self.root, textvariable=self.status_var,
bg="#252526", fg="#858585", font=("Arial", 9),
anchor="w", padx=8).pack(fill=tk.X, side=tk.BOTTOM)
def _on_task_change(self, *_):
task = self.task_var.get()
if task == "分類":
self.algo_cb["values"] = list(self.CLASSIFIERS.keys())
self.algo_var.set("ロジスティック回帰")
elif task == "回帰":
self.algo_cb["values"] = ["線形回帰"]
self.algo_var.set("線形回帰")
else:
self.algo_cb["values"] = ["K-Means"]
self.algo_var.set("K-Means")
def _run(self):
if not SKLEARN_AVAILABLE or not MATPLOTLIB_AVAILABLE:
messagebox.showerror("エラー",
"pip install scikit-learn matplotlib")
return
self.status_var.set("実行中...")
threading.Thread(target=self._do_run, daemon=True).start()
def _do_run(self):
try:
task = self.task_var.get()
n = self.n_samples_var.get()
ts = self.test_size_var.get()
algo = self.algo_var.get()
if task == "分類":
self._run_classification(n, ts, algo)
elif task == "回帰":
self._run_regression(n, ts)
else:
self._run_clustering(n)
except Exception as e:
self.root.after(0, self.status_var.set, f"エラー: {e}")
def _run_classification(self, n, ts, algo_name):
X, y = make_classification(
n_samples=n, n_features=2, n_informative=2,
n_redundant=0, n_clusters_per_class=1, random_state=42)
scaler = StandardScaler()
X_s = scaler.fit_transform(X)
X_tr, X_te, y_tr, y_te = train_test_split(X_s, y, test_size=ts, random_state=42)
models = {
"ロジスティック回帰": LogisticRegression(max_iter=200),
"決定木": DecisionTreeClassifier(max_depth=5),
"ランダムフォレスト": RandomForestClassifier(n_estimators=100),
"SVM": SVC(kernel="rbf", probability=True),
"k近傍法": KNeighborsClassifier(n_neighbors=5),
}
model = models.get(algo_name, LogisticRegression())
model.fit(X_tr, y_tr)
y_pred = model.predict(X_te)
acc = accuracy_score(y_te, y_pred)
report = classification_report(y_te, y_pred, target_names=["クラス0", "クラス1"])
self.root.after(0, self._render_classification,
X_s, y, model, acc, report, algo_name)
def _render_classification(self, X, y, model, acc, report, algo_name):
self.ax.clear()
self.ax.set_facecolor("#0d1117")
# 決定境界
h = 0.05
x_min, x_max = X[:, 0].min() - 0.5, X[:, 0].max() + 0.5
y_min, y_max = X[:, 1].min() - 0.5, X[:, 1].max() + 0.5
xx, yy = np.meshgrid(np.arange(x_min, x_max, h),
np.arange(y_min, y_max, h))
Z = model.predict(np.c_[xx.ravel(), yy.ravel()])
Z = Z.reshape(xx.shape)
self.ax.contourf(xx, yy, Z, alpha=0.25,
cmap="RdYlGn", levels=[-0.5, 0.5, 1.5])
colors = ["#ef5350", "#26a69a"]
for cls in [0, 1]:
mask = y == cls
self.ax.scatter(X[mask, 0], X[mask, 1], color=colors[cls],
s=18, alpha=0.7, label=f"クラス{cls}")
self.ax.set_title(f"{algo_name} 精度: {acc:.3f}", color="#c9d1d9")
self.ax.tick_params(colors="#8b949e", labelsize=7)
for spine in self.ax.spines.values():
spine.set_edgecolor("#30363d")
self.ax.legend(facecolor="#161b22", labelcolor="#c9d1d9",
edgecolor="#30363d", fontsize=8)
self.fig.tight_layout()
self.mpl_canvas.draw()
self._set_report(f"アルゴリズム: {algo_name}\n精度: {acc:.4f}\n\n{report}")
self.status_var.set(f"完了: {algo_name} 精度 {acc:.4f}")
def _run_regression(self, n, ts):
X, y = make_regression(n_samples=n, n_features=1, noise=20, random_state=42)
X_tr, X_te, y_tr, y_te = train_test_split(X, y, test_size=ts, random_state=42)
model = LinearRegression()
model.fit(X_tr, y_tr)
y_pred = model.predict(X_te)
mse = mean_squared_error(y_te, y_pred)
r2 = r2_score(y_te, y_pred)
self.root.after(0, self._render_regression,
X, y, model, mse, r2)
def _render_regression(self, X, y, model, mse, r2):
self.ax.clear()
self.ax.set_facecolor("#0d1117")
self.ax.scatter(X, y, color="#4fc3f7", s=15, alpha=0.6, label="データ")
x_line = np.linspace(X.min(), X.max(), 100).reshape(-1, 1)
y_line = model.predict(x_line)
self.ax.plot(x_line, y_line, color="#ef5350", linewidth=2, label="予測")
self.ax.set_title(f"線形回帰 R²: {r2:.3f}", color="#c9d1d9")
self.ax.tick_params(colors="#8b949e", labelsize=7)
for spine in self.ax.spines.values():
spine.set_edgecolor("#30363d")
self.ax.legend(facecolor="#161b22", labelcolor="#c9d1d9",
edgecolor="#30363d", fontsize=8)
self.fig.tight_layout()
self.mpl_canvas.draw()
self._set_report(
f"アルゴリズム: 線形回帰\n"
f"MSE: {mse:.4f}\nRMSE: {math.sqrt(mse):.4f}\nR²: {r2:.4f}\n"
f"係数: {model.coef_[0]:.4f}\n切片: {model.intercept_:.4f}")
self.status_var.set(f"完了: 線形回帰 R² {r2:.4f}")
def _run_clustering(self, n):
from sklearn.cluster import KMeans
X, _ = make_blobs(n_samples=n, centers=4, random_state=42)
kmeans = KMeans(n_clusters=4, random_state=42, n_init=10)
labels = kmeans.fit_predict(X)
self.root.after(0, self._render_clustering, X, labels, kmeans)
def _render_clustering(self, X, labels, kmeans):
self.ax.clear()
self.ax.set_facecolor("#0d1117")
colors = ["#4fc3f7", "#ef5350", "#26a69a", "#ffa726"]
for i in range(4):
mask = labels == i
self.ax.scatter(X[mask, 0], X[mask, 1], color=colors[i],
s=15, alpha=0.6, label=f"クラスタ{i}")
centers = kmeans.cluster_centers_
self.ax.scatter(centers[:, 0], centers[:, 1],
marker="*", s=200, color="white", zorder=5)
self.ax.set_title("K-Means クラスタリング (k=4)", color="#c9d1d9")
self.ax.tick_params(colors="#8b949e", labelsize=7)
for spine in self.ax.spines.values():
spine.set_edgecolor("#30363d")
self.ax.legend(facecolor="#161b22", labelcolor="#c9d1d9",
edgecolor="#30363d", fontsize=8)
self.fig.tight_layout()
self.mpl_canvas.draw()
inertia = kmeans.inertia_
self._set_report(
f"アルゴリズム: K-Means\nクラスタ数: 4\n"
f"慣性 (inertia): {inertia:.2f}\nサンプル数: {len(X)}")
self.status_var.set(f"完了: K-Means 慣性 {inertia:.2f}")
def _set_report(self, text):
self.report_text.configure(state=tk.NORMAL)
self.report_text.delete("1.0", tk.END)
self.report_text.insert("1.0", text)
self.report_text.configure(state=tk.DISABLED)
if __name__ == "__main__":
root = tk.Tk()
app = App079(root)
root.mainloop()
イベント処理
ボタンのcommand引数でクリックイベントを、bind('
import tkinter as tk
from tkinter import ttk, messagebox
import threading
import random
import math
try:
import matplotlib
matplotlib.use("TkAgg")
from matplotlib.figure import Figure
from matplotlib.backends.backend_tkagg import FigureCanvasTkAgg
MATPLOTLIB_AVAILABLE = True
except ImportError:
MATPLOTLIB_AVAILABLE = False
try:
from sklearn.datasets import make_classification, make_regression, make_blobs
from sklearn.model_selection import train_test_split
from sklearn.preprocessing import StandardScaler
from sklearn.linear_model import LogisticRegression, LinearRegression
from sklearn.tree import DecisionTreeClassifier
from sklearn.ensemble import RandomForestClassifier
from sklearn.svm import SVC
from sklearn.neighbors import KNeighborsClassifier
from sklearn.metrics import (accuracy_score, classification_report,
mean_squared_error, r2_score)
import numpy as np
SKLEARN_AVAILABLE = True
except ImportError:
SKLEARN_AVAILABLE = False
class App079:
"""機械学習デモアプリ"""
CLASSIFIERS = {
"ロジスティック回帰": "LogisticRegression",
"決定木": "DecisionTreeClassifier",
"ランダムフォレスト": "RandomForestClassifier",
"SVM": "SVC",
"k近傍法": "KNeighborsClassifier",
}
def __init__(self, root):
self.root = root
self.root.title("機械学習デモアプリ")
self.root.geometry("1000x640")
self.root.configure(bg="#1e1e1e")
self._build_ui()
def _build_ui(self):
header = tk.Frame(self.root, bg="#252526", pady=6)
header.pack(fill=tk.X)
tk.Label(header, text="🤖 機械学習デモアプリ",
font=("Noto Sans JP", 12, "bold"),
bg="#252526", fg="#4fc3f7").pack(side=tk.LEFT, padx=12)
missing = []
if not SKLEARN_AVAILABLE: missing.append("scikit-learn")
if not MATPLOTLIB_AVAILABLE: missing.append("matplotlib")
if missing:
tk.Label(self.root,
text=f"⚠ pip install {' '.join(missing)}",
bg="#fff3cd", fg="#856404", font=("Arial", 9),
anchor="w", padx=8).pack(fill=tk.X)
# 設定パネル
cfg = tk.LabelFrame(self.root, text="設定", bg="#252526",
fg="#ccc", font=("Arial", 9), padx=8, pady=6)
cfg.pack(fill=tk.X, padx=8, pady=4)
# タスク
tk.Label(cfg, text="タスク:", bg="#252526", fg="#ccc",
font=("Arial", 9)).grid(row=0, column=0, sticky="w")
self.task_var = tk.StringVar(value="分類")
ttk.Combobox(cfg, textvariable=self.task_var,
values=["分類", "回帰", "クラスタリング"],
state="readonly", width=12).grid(row=0, column=1, padx=4)
# アルゴリズム
tk.Label(cfg, text="アルゴリズム:", bg="#252526", fg="#ccc",
font=("Arial", 9)).grid(row=0, column=2, sticky="w", padx=(12, 0))
self.algo_var = tk.StringVar(value="ロジスティック回帰")
self.algo_cb = ttk.Combobox(cfg, textvariable=self.algo_var,
values=list(self.CLASSIFIERS.keys()),
state="readonly", width=18)
self.algo_cb.grid(row=0, column=3, padx=4)
# サンプル数
tk.Label(cfg, text="サンプル数:", bg="#252526", fg="#ccc",
font=("Arial", 9)).grid(row=0, column=4, sticky="w", padx=(12, 0))
self.n_samples_var = tk.IntVar(value=300)
ttk.Spinbox(cfg, from_=50, to=2000, textvariable=self.n_samples_var,
width=6).grid(row=0, column=5, padx=4)
# テスト割合
tk.Label(cfg, text="テスト割合:", bg="#252526", fg="#ccc",
font=("Arial", 9)).grid(row=0, column=6, sticky="w", padx=(8, 0))
self.test_size_var = tk.DoubleVar(value=0.2)
ttk.Spinbox(cfg, from_=0.1, to=0.5, increment=0.05,
textvariable=self.test_size_var,
width=5).grid(row=0, column=7, padx=4)
tk.Button(cfg, text="▶ 実行", command=self._run,
bg="#1565c0", fg="white", relief=tk.FLAT,
font=("Arial", 10, "bold"), padx=14, pady=4,
activebackground="#0d47a1", bd=0).grid(
row=0, column=8, padx=8)
self.task_var.trace_add("write", self._on_task_change)
# メイン: 左=グラフ, 右=レポート
main = tk.Frame(self.root, bg="#1e1e1e")
main.pack(fill=tk.BOTH, expand=True, padx=8, pady=4)
# 左: グラフ
left = tk.Frame(main, bg="#0d1117")
left.pack(side=tk.LEFT, fill=tk.BOTH, expand=True)
if MATPLOTLIB_AVAILABLE:
self.fig = Figure(figsize=(6, 5), facecolor="#0d1117")
self.ax = self.fig.add_subplot(111)
self.mpl_canvas = FigureCanvasTkAgg(self.fig, master=left)
self.mpl_canvas.get_tk_widget().pack(fill=tk.BOTH, expand=True)
# 右: 結果レポート
right = tk.Frame(main, bg="#1e1e1e", width=340)
right.pack(side=tk.LEFT, fill=tk.Y, padx=(8, 0))
right.pack_propagate(False)
tk.Label(right, text="評価レポート", bg="#1e1e1e", fg="#888",
font=("Arial", 9)).pack(anchor="w")
self.report_text = tk.Text(right, bg="#0d1117", fg="#c9d1d9",
font=("Courier New", 9), relief=tk.FLAT,
state=tk.DISABLED)
self.report_text.pack(fill=tk.BOTH, expand=True)
self.status_var = tk.StringVar(value="「実行」ボタンで機械学習デモを開始します")
tk.Label(self.root, textvariable=self.status_var,
bg="#252526", fg="#858585", font=("Arial", 9),
anchor="w", padx=8).pack(fill=tk.X, side=tk.BOTTOM)
def _on_task_change(self, *_):
task = self.task_var.get()
if task == "分類":
self.algo_cb["values"] = list(self.CLASSIFIERS.keys())
self.algo_var.set("ロジスティック回帰")
elif task == "回帰":
self.algo_cb["values"] = ["線形回帰"]
self.algo_var.set("線形回帰")
else:
self.algo_cb["values"] = ["K-Means"]
self.algo_var.set("K-Means")
def _run(self):
if not SKLEARN_AVAILABLE or not MATPLOTLIB_AVAILABLE:
messagebox.showerror("エラー",
"pip install scikit-learn matplotlib")
return
self.status_var.set("実行中...")
threading.Thread(target=self._do_run, daemon=True).start()
def _do_run(self):
try:
task = self.task_var.get()
n = self.n_samples_var.get()
ts = self.test_size_var.get()
algo = self.algo_var.get()
if task == "分類":
self._run_classification(n, ts, algo)
elif task == "回帰":
self._run_regression(n, ts)
else:
self._run_clustering(n)
except Exception as e:
self.root.after(0, self.status_var.set, f"エラー: {e}")
def _run_classification(self, n, ts, algo_name):
X, y = make_classification(
n_samples=n, n_features=2, n_informative=2,
n_redundant=0, n_clusters_per_class=1, random_state=42)
scaler = StandardScaler()
X_s = scaler.fit_transform(X)
X_tr, X_te, y_tr, y_te = train_test_split(X_s, y, test_size=ts, random_state=42)
models = {
"ロジスティック回帰": LogisticRegression(max_iter=200),
"決定木": DecisionTreeClassifier(max_depth=5),
"ランダムフォレスト": RandomForestClassifier(n_estimators=100),
"SVM": SVC(kernel="rbf", probability=True),
"k近傍法": KNeighborsClassifier(n_neighbors=5),
}
model = models.get(algo_name, LogisticRegression())
model.fit(X_tr, y_tr)
y_pred = model.predict(X_te)
acc = accuracy_score(y_te, y_pred)
report = classification_report(y_te, y_pred, target_names=["クラス0", "クラス1"])
self.root.after(0, self._render_classification,
X_s, y, model, acc, report, algo_name)
def _render_classification(self, X, y, model, acc, report, algo_name):
self.ax.clear()
self.ax.set_facecolor("#0d1117")
# 決定境界
h = 0.05
x_min, x_max = X[:, 0].min() - 0.5, X[:, 0].max() + 0.5
y_min, y_max = X[:, 1].min() - 0.5, X[:, 1].max() + 0.5
xx, yy = np.meshgrid(np.arange(x_min, x_max, h),
np.arange(y_min, y_max, h))
Z = model.predict(np.c_[xx.ravel(), yy.ravel()])
Z = Z.reshape(xx.shape)
self.ax.contourf(xx, yy, Z, alpha=0.25,
cmap="RdYlGn", levels=[-0.5, 0.5, 1.5])
colors = ["#ef5350", "#26a69a"]
for cls in [0, 1]:
mask = y == cls
self.ax.scatter(X[mask, 0], X[mask, 1], color=colors[cls],
s=18, alpha=0.7, label=f"クラス{cls}")
self.ax.set_title(f"{algo_name} 精度: {acc:.3f}", color="#c9d1d9")
self.ax.tick_params(colors="#8b949e", labelsize=7)
for spine in self.ax.spines.values():
spine.set_edgecolor("#30363d")
self.ax.legend(facecolor="#161b22", labelcolor="#c9d1d9",
edgecolor="#30363d", fontsize=8)
self.fig.tight_layout()
self.mpl_canvas.draw()
self._set_report(f"アルゴリズム: {algo_name}\n精度: {acc:.4f}\n\n{report}")
self.status_var.set(f"完了: {algo_name} 精度 {acc:.4f}")
def _run_regression(self, n, ts):
X, y = make_regression(n_samples=n, n_features=1, noise=20, random_state=42)
X_tr, X_te, y_tr, y_te = train_test_split(X, y, test_size=ts, random_state=42)
model = LinearRegression()
model.fit(X_tr, y_tr)
y_pred = model.predict(X_te)
mse = mean_squared_error(y_te, y_pred)
r2 = r2_score(y_te, y_pred)
self.root.after(0, self._render_regression,
X, y, model, mse, r2)
def _render_regression(self, X, y, model, mse, r2):
self.ax.clear()
self.ax.set_facecolor("#0d1117")
self.ax.scatter(X, y, color="#4fc3f7", s=15, alpha=0.6, label="データ")
x_line = np.linspace(X.min(), X.max(), 100).reshape(-1, 1)
y_line = model.predict(x_line)
self.ax.plot(x_line, y_line, color="#ef5350", linewidth=2, label="予測")
self.ax.set_title(f"線形回帰 R²: {r2:.3f}", color="#c9d1d9")
self.ax.tick_params(colors="#8b949e", labelsize=7)
for spine in self.ax.spines.values():
spine.set_edgecolor("#30363d")
self.ax.legend(facecolor="#161b22", labelcolor="#c9d1d9",
edgecolor="#30363d", fontsize=8)
self.fig.tight_layout()
self.mpl_canvas.draw()
self._set_report(
f"アルゴリズム: 線形回帰\n"
f"MSE: {mse:.4f}\nRMSE: {math.sqrt(mse):.4f}\nR²: {r2:.4f}\n"
f"係数: {model.coef_[0]:.4f}\n切片: {model.intercept_:.4f}")
self.status_var.set(f"完了: 線形回帰 R² {r2:.4f}")
def _run_clustering(self, n):
from sklearn.cluster import KMeans
X, _ = make_blobs(n_samples=n, centers=4, random_state=42)
kmeans = KMeans(n_clusters=4, random_state=42, n_init=10)
labels = kmeans.fit_predict(X)
self.root.after(0, self._render_clustering, X, labels, kmeans)
def _render_clustering(self, X, labels, kmeans):
self.ax.clear()
self.ax.set_facecolor("#0d1117")
colors = ["#4fc3f7", "#ef5350", "#26a69a", "#ffa726"]
for i in range(4):
mask = labels == i
self.ax.scatter(X[mask, 0], X[mask, 1], color=colors[i],
s=15, alpha=0.6, label=f"クラスタ{i}")
centers = kmeans.cluster_centers_
self.ax.scatter(centers[:, 0], centers[:, 1],
marker="*", s=200, color="white", zorder=5)
self.ax.set_title("K-Means クラスタリング (k=4)", color="#c9d1d9")
self.ax.tick_params(colors="#8b949e", labelsize=7)
for spine in self.ax.spines.values():
spine.set_edgecolor("#30363d")
self.ax.legend(facecolor="#161b22", labelcolor="#c9d1d9",
edgecolor="#30363d", fontsize=8)
self.fig.tight_layout()
self.mpl_canvas.draw()
inertia = kmeans.inertia_
self._set_report(
f"アルゴリズム: K-Means\nクラスタ数: 4\n"
f"慣性 (inertia): {inertia:.2f}\nサンプル数: {len(X)}")
self.status_var.set(f"完了: K-Means 慣性 {inertia:.2f}")
def _set_report(self, text):
self.report_text.configure(state=tk.NORMAL)
self.report_text.delete("1.0", tk.END)
self.report_text.insert("1.0", text)
self.report_text.configure(state=tk.DISABLED)
if __name__ == "__main__":
root = tk.Tk()
app = App079(root)
root.mainloop()
Textウィジェットでの結果表示
tk.Textウィジェットをstate=DISABLED（読み取り専用）で作成し、更新時はNORMALに変更してinsert()で内容を書き込み、再びDISABLEDに戻します。
import tkinter as tk
from tkinter import ttk, messagebox
import threading
import random
import math
try:
import matplotlib
matplotlib.use("TkAgg")
from matplotlib.figure import Figure
from matplotlib.backends.backend_tkagg import FigureCanvasTkAgg
MATPLOTLIB_AVAILABLE = True
except ImportError:
MATPLOTLIB_AVAILABLE = False
try:
from sklearn.datasets import make_classification, make_regression, make_blobs
from sklearn.model_selection import train_test_split
from sklearn.preprocessing import StandardScaler
from sklearn.linear_model import LogisticRegression, LinearRegression
from sklearn.tree import DecisionTreeClassifier
from sklearn.ensemble import RandomForestClassifier
from sklearn.svm import SVC
from sklearn.neighbors import KNeighborsClassifier
from sklearn.metrics import (accuracy_score, classification_report,
mean_squared_error, r2_score)
import numpy as np
SKLEARN_AVAILABLE = True
except ImportError:
SKLEARN_AVAILABLE = False
class App079:
"""機械学習デモアプリ"""
CLASSIFIERS = {
"ロジスティック回帰": "LogisticRegression",
"決定木": "DecisionTreeClassifier",
"ランダムフォレスト": "RandomForestClassifier",
"SVM": "SVC",
"k近傍法": "KNeighborsClassifier",
}
def __init__(self, root):
self.root = root
self.root.title("機械学習デモアプリ")
self.root.geometry("1000x640")
self.root.configure(bg="#1e1e1e")
self._build_ui()
def _build_ui(self):
header = tk.Frame(self.root, bg="#252526", pady=6)
header.pack(fill=tk.X)
tk.Label(header, text="🤖 機械学習デモアプリ",
font=("Noto Sans JP", 12, "bold"),
bg="#252526", fg="#4fc3f7").pack(side=tk.LEFT, padx=12)
missing = []
if not SKLEARN_AVAILABLE: missing.append("scikit-learn")
if not MATPLOTLIB_AVAILABLE: missing.append("matplotlib")
if missing:
tk.Label(self.root,
text=f"⚠ pip install {' '.join(missing)}",
bg="#fff3cd", fg="#856404", font=("Arial", 9),
anchor="w", padx=8).pack(fill=tk.X)
# 設定パネル
cfg = tk.LabelFrame(self.root, text="設定", bg="#252526",
fg="#ccc", font=("Arial", 9), padx=8, pady=6)
cfg.pack(fill=tk.X, padx=8, pady=4)
# タスク
tk.Label(cfg, text="タスク:", bg="#252526", fg="#ccc",
font=("Arial", 9)).grid(row=0, column=0, sticky="w")
self.task_var = tk.StringVar(value="分類")
ttk.Combobox(cfg, textvariable=self.task_var,
values=["分類", "回帰", "クラスタリング"],
state="readonly", width=12).grid(row=0, column=1, padx=4)
# アルゴリズム
tk.Label(cfg, text="アルゴリズム:", bg="#252526", fg="#ccc",
font=("Arial", 9)).grid(row=0, column=2, sticky="w", padx=(12, 0))
self.algo_var = tk.StringVar(value="ロジスティック回帰")
self.algo_cb = ttk.Combobox(cfg, textvariable=self.algo_var,
values=list(self.CLASSIFIERS.keys()),
state="readonly", width=18)
self.algo_cb.grid(row=0, column=3, padx=4)
# サンプル数
tk.Label(cfg, text="サンプル数:", bg="#252526", fg="#ccc",
font=("Arial", 9)).grid(row=0, column=4, sticky="w", padx=(12, 0))
self.n_samples_var = tk.IntVar(value=300)
ttk.Spinbox(cfg, from_=50, to=2000, textvariable=self.n_samples_var,
width=6).grid(row=0, column=5, padx=4)
# テスト割合
tk.Label(cfg, text="テスト割合:", bg="#252526", fg="#ccc",
font=("Arial", 9)).grid(row=0, column=6, sticky="w", padx=(8, 0))
self.test_size_var = tk.DoubleVar(value=0.2)
ttk.Spinbox(cfg, from_=0.1, to=0.5, increment=0.05,
textvariable=self.test_size_var,
width=5).grid(row=0, column=7, padx=4)
tk.Button(cfg, text="▶ 実行", command=self._run,
bg="#1565c0", fg="white", relief=tk.FLAT,
font=("Arial", 10, "bold"), padx=14, pady=4,
activebackground="#0d47a1", bd=0).grid(
row=0, column=8, padx=8)
self.task_var.trace_add("write", self._on_task_change)
# メイン: 左=グラフ, 右=レポート
main = tk.Frame(self.root, bg="#1e1e1e")
main.pack(fill=tk.BOTH, expand=True, padx=8, pady=4)
# 左: グラフ
left = tk.Frame(main, bg="#0d1117")
left.pack(side=tk.LEFT, fill=tk.BOTH, expand=True)
if MATPLOTLIB_AVAILABLE:
self.fig = Figure(figsize=(6, 5), facecolor="#0d1117")
self.ax = self.fig.add_subplot(111)
self.mpl_canvas = FigureCanvasTkAgg(self.fig, master=left)
self.mpl_canvas.get_tk_widget().pack(fill=tk.BOTH, expand=True)
# 右: 結果レポート
right = tk.Frame(main, bg="#1e1e1e", width=340)
right.pack(side=tk.LEFT, fill=tk.Y, padx=(8, 0))
right.pack_propagate(False)
tk.Label(right, text="評価レポート", bg="#1e1e1e", fg="#888",
font=("Arial", 9)).pack(anchor="w")
self.report_text = tk.Text(right, bg="#0d1117", fg="#c9d1d9",
font=("Courier New", 9), relief=tk.FLAT,
state=tk.DISABLED)
self.report_text.pack(fill=tk.BOTH, expand=True)
self.status_var = tk.StringVar(value="「実行」ボタンで機械学習デモを開始します")
tk.Label(self.root, textvariable=self.status_var,
bg="#252526", fg="#858585", font=("Arial", 9),
anchor="w", padx=8).pack(fill=tk.X, side=tk.BOTTOM)
def _on_task_change(self, *_):
task = self.task_var.get()
if task == "分類":
self.algo_cb["values"] = list(self.CLASSIFIERS.keys())
self.algo_var.set("ロジスティック回帰")
elif task == "回帰":
self.algo_cb["values"] = ["線形回帰"]
self.algo_var.set("線形回帰")
else:
self.algo_cb["values"] = ["K-Means"]
self.algo_var.set("K-Means")
def _run(self):
if not SKLEARN_AVAILABLE or not MATPLOTLIB_AVAILABLE:
messagebox.showerror("エラー",
"pip install scikit-learn matplotlib")
return
self.status_var.set("実行中...")
threading.Thread(target=self._do_run, daemon=True).start()
def _do_run(self):
try:
task = self.task_var.get()
n = self.n_samples_var.get()
ts = self.test_size_var.get()
algo = self.algo_var.get()
if task == "分類":
self._run_classification(n, ts, algo)
elif task == "回帰":
self._run_regression(n, ts)
else:
self._run_clustering(n)
except Exception as e:
self.root.after(0, self.status_var.set, f"エラー: {e}")
def _run_classification(self, n, ts, algo_name):
X, y = make_classification(
n_samples=n, n_features=2, n_informative=2,
n_redundant=0, n_clusters_per_class=1, random_state=42)
scaler = StandardScaler()
X_s = scaler.fit_transform(X)
X_tr, X_te, y_tr, y_te = train_test_split(X_s, y, test_size=ts, random_state=42)
models = {
"ロジスティック回帰": LogisticRegression(max_iter=200),
"決定木": DecisionTreeClassifier(max_depth=5),
"ランダムフォレスト": RandomForestClassifier(n_estimators=100),
"SVM": SVC(kernel="rbf", probability=True),
"k近傍法": KNeighborsClassifier(n_neighbors=5),
}
model = models.get(algo_name, LogisticRegression())
model.fit(X_tr, y_tr)
y_pred = model.predict(X_te)
acc = accuracy_score(y_te, y_pred)
report = classification_report(y_te, y_pred, target_names=["クラス0", "クラス1"])
self.root.after(0, self._render_classification,
X_s, y, model, acc, report, algo_name)
def _render_classification(self, X, y, model, acc, report, algo_name):
self.ax.clear()
self.ax.set_facecolor("#0d1117")
# 決定境界
h = 0.05
x_min, x_max = X[:, 0].min() - 0.5, X[:, 0].max() + 0.5
y_min, y_max = X[:, 1].min() - 0.5, X[:, 1].max() + 0.5
xx, yy = np.meshgrid(np.arange(x_min, x_max, h),
np.arange(y_min, y_max, h))
Z = model.predict(np.c_[xx.ravel(), yy.ravel()])
Z = Z.reshape(xx.shape)
self.ax.contourf(xx, yy, Z, alpha=0.25,
cmap="RdYlGn", levels=[-0.5, 0.5, 1.5])
colors = ["#ef5350", "#26a69a"]
for cls in [0, 1]:
mask = y == cls
self.ax.scatter(X[mask, 0], X[mask, 1], color=colors[cls],
s=18, alpha=0.7, label=f"クラス{cls}")
self.ax.set_title(f"{algo_name} 精度: {acc:.3f}", color="#c9d1d9")
self.ax.tick_params(colors="#8b949e", labelsize=7)
for spine in self.ax.spines.values():
spine.set_edgecolor("#30363d")
self.ax.legend(facecolor="#161b22", labelcolor="#c9d1d9",
edgecolor="#30363d", fontsize=8)
self.fig.tight_layout()
self.mpl_canvas.draw()
self._set_report(f"アルゴリズム: {algo_name}\n精度: {acc:.4f}\n\n{report}")
self.status_var.set(f"完了: {algo_name} 精度 {acc:.4f}")
def _run_regression(self, n, ts):
X, y = make_regression(n_samples=n, n_features=1, noise=20, random_state=42)
X_tr, X_te, y_tr, y_te = train_test_split(X, y, test_size=ts, random_state=42)
model = LinearRegression()
model.fit(X_tr, y_tr)
y_pred = model.predict(X_te)
mse = mean_squared_error(y_te, y_pred)
r2 = r2_score(y_te, y_pred)
self.root.after(0, self._render_regression,
X, y, model, mse, r2)
def _render_regression(self, X, y, model, mse, r2):
self.ax.clear()
self.ax.set_facecolor("#0d1117")
self.ax.scatter(X, y, color="#4fc3f7", s=15, alpha=0.6, label="データ")
x_line = np.linspace(X.min(), X.max(), 100).reshape(-1, 1)
y_line = model.predict(x_line)
self.ax.plot(x_line, y_line, color="#ef5350", linewidth=2, label="予測")
self.ax.set_title(f"線形回帰 R²: {r2:.3f}", color="#c9d1d9")
self.ax.tick_params(colors="#8b949e", labelsize=7)
for spine in self.ax.spines.values():
spine.set_edgecolor("#30363d")
self.ax.legend(facecolor="#161b22", labelcolor="#c9d1d9",
edgecolor="#30363d", fontsize=8)
self.fig.tight_layout()
self.mpl_canvas.draw()
self._set_report(
f"アルゴリズム: 線形回帰\n"
f"MSE: {mse:.4f}\nRMSE: {math.sqrt(mse):.4f}\nR²: {r2:.4f}\n"
f"係数: {model.coef_[0]:.4f}\n切片: {model.intercept_:.4f}")
self.status_var.set(f"完了: 線形回帰 R² {r2:.4f}")
def _run_clustering(self, n):
from sklearn.cluster import KMeans
X, _ = make_blobs(n_samples=n, centers=4, random_state=42)
kmeans = KMeans(n_clusters=4, random_state=42, n_init=10)
labels = kmeans.fit_predict(X)
self.root.after(0, self._render_clustering, X, labels, kmeans)
def _render_clustering(self, X, labels, kmeans):
self.ax.clear()
self.ax.set_facecolor("#0d1117")
colors = ["#4fc3f7", "#ef5350", "#26a69a", "#ffa726"]
for i in range(4):
mask = labels == i
self.ax.scatter(X[mask, 0], X[mask, 1], color=colors[i],
s=15, alpha=0.6, label=f"クラスタ{i}")
centers = kmeans.cluster_centers_
self.ax.scatter(centers[:, 0], centers[:, 1],
marker="*", s=200, color="white", zorder=5)
self.ax.set_title("K-Means クラスタリング (k=4)", color="#c9d1d9")
self.ax.tick_params(colors="#8b949e", labelsize=7)
for spine in self.ax.spines.values():
spine.set_edgecolor("#30363d")
self.ax.legend(facecolor="#161b22", labelcolor="#c9d1d9",
edgecolor="#30363d", fontsize=8)
self.fig.tight_layout()
self.mpl_canvas.draw()
inertia = kmeans.inertia_
self._set_report(
f"アルゴリズム: K-Means\nクラスタ数: 4\n"
f"慣性 (inertia): {inertia:.2f}\nサンプル数: {len(X)}")
self.status_var.set(f"完了: K-Means 慣性 {inertia:.2f}")
def _set_report(self, text):
self.report_text.configure(state=tk.NORMAL)
self.report_text.delete("1.0", tk.END)
self.report_text.insert("1.0", text)
self.report_text.configure(state=tk.DISABLED)
if __name__ == "__main__":
root = tk.Tk()
app = App079(root)
root.mainloop()
例外処理とエラーハンドリング
try-exceptでValueErrorとExceptionを捕捉し、messagebox.showerror()でエラーメッセージを表示します。予期しないエラーも処理することで、アプリの堅牢性が向上します。
import tkinter as tk
from tkinter import ttk, messagebox
import threading
import random
import math
try:
import matplotlib
matplotlib.use("TkAgg")
from matplotlib.figure import Figure
from matplotlib.backends.backend_tkagg import FigureCanvasTkAgg
MATPLOTLIB_AVAILABLE = True
except ImportError:
MATPLOTLIB_AVAILABLE = False
try:
from sklearn.datasets import make_classification, make_regression, make_blobs
from sklearn.model_selection import train_test_split
from sklearn.preprocessing import StandardScaler
from sklearn.linear_model import LogisticRegression, LinearRegression
from sklearn.tree import DecisionTreeClassifier
from sklearn.ensemble import RandomForestClassifier
from sklearn.svm import SVC
from sklearn.neighbors import KNeighborsClassifier
from sklearn.metrics import (accuracy_score, classification_report,
mean_squared_error, r2_score)
import numpy as np
SKLEARN_AVAILABLE = True
except ImportError:
SKLEARN_AVAILABLE = False
class App079:
"""機械学習デモアプリ"""
CLASSIFIERS = {
"ロジスティック回帰": "LogisticRegression",
"決定木": "DecisionTreeClassifier",
"ランダムフォレスト": "RandomForestClassifier",
"SVM": "SVC",
"k近傍法": "KNeighborsClassifier",
}
def __init__(self, root):
self.root = root
self.root.title("機械学習デモアプリ")
self.root.geometry("1000x640")
self.root.configure(bg="#1e1e1e")
self._build_ui()
def _build_ui(self):
header = tk.Frame(self.root, bg="#252526", pady=6)
header.pack(fill=tk.X)
tk.Label(header, text="🤖 機械学習デモアプリ",
font=("Noto Sans JP", 12, "bold"),
bg="#252526", fg="#4fc3f7").pack(side=tk.LEFT, padx=12)
missing = []
if not SKLEARN_AVAILABLE: missing.append("scikit-learn")
if not MATPLOTLIB_AVAILABLE: missing.append("matplotlib")
if missing:
tk.Label(self.root,
text=f"⚠ pip install {' '.join(missing)}",
bg="#fff3cd", fg="#856404", font=("Arial", 9),
anchor="w", padx=8).pack(fill=tk.X)
# 設定パネル
cfg = tk.LabelFrame(self.root, text="設定", bg="#252526",
fg="#ccc", font=("Arial", 9), padx=8, pady=6)
cfg.pack(fill=tk.X, padx=8, pady=4)
# タスク
tk.Label(cfg, text="タスク:", bg="#252526", fg="#ccc",
font=("Arial", 9)).grid(row=0, column=0, sticky="w")
self.task_var = tk.StringVar(value="分類")
ttk.Combobox(cfg, textvariable=self.task_var,
values=["分類", "回帰", "クラスタリング"],
state="readonly", width=12).grid(row=0, column=1, padx=4)
# アルゴリズム
tk.Label(cfg, text="アルゴリズム:", bg="#252526", fg="#ccc",
font=("Arial", 9)).grid(row=0, column=2, sticky="w", padx=(12, 0))
self.algo_var = tk.StringVar(value="ロジスティック回帰")
self.algo_cb = ttk.Combobox(cfg, textvariable=self.algo_var,
values=list(self.CLASSIFIERS.keys()),
state="readonly", width=18)
self.algo_cb.grid(row=0, column=3, padx=4)
# サンプル数
tk.Label(cfg, text="サンプル数:", bg="#252526", fg="#ccc",
font=("Arial", 9)).grid(row=0, column=4, sticky="w", padx=(12, 0))
self.n_samples_var = tk.IntVar(value=300)
ttk.Spinbox(cfg, from_=50, to=2000, textvariable=self.n_samples_var,
width=6).grid(row=0, column=5, padx=4)
# テスト割合
tk.Label(cfg, text="テスト割合:", bg="#252526", fg="#ccc",
font=("Arial", 9)).grid(row=0, column=6, sticky="w", padx=(8, 0))
self.test_size_var = tk.DoubleVar(value=0.2)
ttk.Spinbox(cfg, from_=0.1, to=0.5, increment=0.05,
textvariable=self.test_size_var,
width=5).grid(row=0, column=7, padx=4)
tk.Button(cfg, text="▶ 実行", command=self._run,
bg="#1565c0", fg="white", relief=tk.FLAT,
font=("Arial", 10, "bold"), padx=14, pady=4,
activebackground="#0d47a1", bd=0).grid(
row=0, column=8, padx=8)
self.task_var.trace_add("write", self._on_task_change)
# メイン: 左=グラフ, 右=レポート
main = tk.Frame(self.root, bg="#1e1e1e")
main.pack(fill=tk.BOTH, expand=True, padx=8, pady=4)
# 左: グラフ
left = tk.Frame(main, bg="#0d1117")
left.pack(side=tk.LEFT, fill=tk.BOTH, expand=True)
if MATPLOTLIB_AVAILABLE:
self.fig = Figure(figsize=(6, 5), facecolor="#0d1117")
self.ax = self.fig.add_subplot(111)
self.mpl_canvas = FigureCanvasTkAgg(self.fig, master=left)
self.mpl_canvas.get_tk_widget().pack(fill=tk.BOTH, expand=True)
# 右: 結果レポート
right = tk.Frame(main, bg="#1e1e1e", width=340)
right.pack(side=tk.LEFT, fill=tk.Y, padx=(8, 0))
right.pack_propagate(False)
tk.Label(right, text="評価レポート", bg="#1e1e1e", fg="#888",
font=("Arial", 9)).pack(anchor="w")
self.report_text = tk.Text(right, bg="#0d1117", fg="#c9d1d9",
font=("Courier New", 9), relief=tk.FLAT,
state=tk.DISABLED)
self.report_text.pack(fill=tk.BOTH, expand=True)
self.status_var = tk.StringVar(value="「実行」ボタンで機械学習デモを開始します")
tk.Label(self.root, textvariable=self.status_var,
bg="#252526", fg="#858585", font=("Arial", 9),
anchor="w", padx=8).pack(fill=tk.X, side=tk.BOTTOM)
def _on_task_change(self, *_):
task = self.task_var.get()
if task == "分類":
self.algo_cb["values"] = list(self.CLASSIFIERS.keys())
self.algo_var.set("ロジスティック回帰")
elif task == "回帰":
self.algo_cb["values"] = ["線形回帰"]
self.algo_var.set("線形回帰")
else:
self.algo_cb["values"] = ["K-Means"]
self.algo_var.set("K-Means")
def _run(self):
if not SKLEARN_AVAILABLE or not MATPLOTLIB_AVAILABLE:
messagebox.showerror("エラー",
"pip install scikit-learn matplotlib")
return
self.status_var.set("実行中...")
threading.Thread(target=self._do_run, daemon=True).start()
def _do_run(self):
try:
task = self.task_var.get()
n = self.n_samples_var.get()
ts = self.test_size_var.get()
algo = self.algo_var.get()
if task == "分類":
self._run_classification(n, ts, algo)
elif task == "回帰":
self._run_regression(n, ts)
else:
self._run_clustering(n)
except Exception as e:
self.root.after(0, self.status_var.set, f"エラー: {e}")
def _run_classification(self, n, ts, algo_name):
X, y = make_classification(
n_samples=n, n_features=2, n_informative=2,
n_redundant=0, n_clusters_per_class=1, random_state=42)
scaler = StandardScaler()
X_s = scaler.fit_transform(X)
X_tr, X_te, y_tr, y_te = train_test_split(X_s, y, test_size=ts, random_state=42)
models = {
"ロジスティック回帰": LogisticRegression(max_iter=200),
"決定木": DecisionTreeClassifier(max_depth=5),
"ランダムフォレスト": RandomForestClassifier(n_estimators=100),
"SVM": SVC(kernel="rbf", probability=True),
"k近傍法": KNeighborsClassifier(n_neighbors=5),
}
model = models.get(algo_name, LogisticRegression())
model.fit(X_tr, y_tr)
y_pred = model.predict(X_te)
acc = accuracy_score(y_te, y_pred)
report = classification_report(y_te, y_pred, target_names=["クラス0", "クラス1"])
self.root.after(0, self._render_classification,
X_s, y, model, acc, report, algo_name)
def _render_classification(self, X, y, model, acc, report, algo_name):
self.ax.clear()
self.ax.set_facecolor("#0d1117")
# 決定境界
h = 0.05
x_min, x_max = X[:, 0].min() - 0.5, X[:, 0].max() + 0.5
y_min, y_max = X[:, 1].min() - 0.5, X[:, 1].max() + 0.5
xx, yy = np.meshgrid(np.arange(x_min, x_max, h),
np.arange(y_min, y_max, h))
Z = model.predict(np.c_[xx.ravel(), yy.ravel()])
Z = Z.reshape(xx.shape)
self.ax.contourf(xx, yy, Z, alpha=0.25,
cmap="RdYlGn", levels=[-0.5, 0.5, 1.5])
colors = ["#ef5350", "#26a69a"]
for cls in [0, 1]:
mask = y == cls
self.ax.scatter(X[mask, 0], X[mask, 1], color=colors[cls],
s=18, alpha=0.7, label=f"クラス{cls}")
self.ax.set_title(f"{algo_name} 精度: {acc:.3f}", color="#c9d1d9")
self.ax.tick_params(colors="#8b949e", labelsize=7)
for spine in self.ax.spines.values():
spine.set_edgecolor("#30363d")
self.ax.legend(facecolor="#161b22", labelcolor="#c9d1d9",
edgecolor="#30363d", fontsize=8)
self.fig.tight_layout()
self.mpl_canvas.draw()
self._set_report(f"アルゴリズム: {algo_name}\n精度: {acc:.4f}\n\n{report}")
self.status_var.set(f"完了: {algo_name} 精度 {acc:.4f}")
def _run_regression(self, n, ts):
X, y = make_regression(n_samples=n, n_features=1, noise=20, random_state=42)
X_tr, X_te, y_tr, y_te = train_test_split(X, y, test_size=ts, random_state=42)
model = LinearRegression()
model.fit(X_tr, y_tr)
y_pred = model.predict(X_te)
mse = mean_squared_error(y_te, y_pred)
r2 = r2_score(y_te, y_pred)
self.root.after(0, self._render_regression,
X, y, model, mse, r2)
def _render_regression(self, X, y, model, mse, r2):
self.ax.clear()
self.ax.set_facecolor("#0d1117")
self.ax.scatter(X, y, color="#4fc3f7", s=15, alpha=0.6, label="データ")
x_line = np.linspace(X.min(), X.max(), 100).reshape(-1, 1)
y_line = model.predict(x_line)
self.ax.plot(x_line, y_line, color="#ef5350", linewidth=2, label="予測")
self.ax.set_title(f"線形回帰 R²: {r2:.3f}", color="#c9d1d9")
self.ax.tick_params(colors="#8b949e", labelsize=7)
for spine in self.ax.spines.values():
spine.set_edgecolor("#30363d")
self.ax.legend(facecolor="#161b22", labelcolor="#c9d1d9",
edgecolor="#30363d", fontsize=8)
self.fig.tight_layout()
self.mpl_canvas.draw()
self._set_report(
f"アルゴリズム: 線形回帰\n"
f"MSE: {mse:.4f}\nRMSE: {math.sqrt(mse):.4f}\nR²: {r2:.4f}\n"
f"係数: {model.coef_[0]:.4f}\n切片: {model.intercept_:.4f}")
self.status_var.set(f"完了: 線形回帰 R² {r2:.4f}")
def _run_clustering(self, n):
from sklearn.cluster import KMeans
X, _ = make_blobs(n_samples=n, centers=4, random_state=42)
kmeans = KMeans(n_clusters=4, random_state=42, n_init=10)
labels = kmeans.fit_predict(X)
self.root.after(0, self._render_clustering, X, labels, kmeans)
def _render_clustering(self, X, labels, kmeans):
self.ax.clear()
self.ax.set_facecolor("#0d1117")
colors = ["#4fc3f7", "#ef5350", "#26a69a", "#ffa726"]
for i in range(4):
mask = labels == i
self.ax.scatter(X[mask, 0], X[mask, 1], color=colors[i],
s=15, alpha=0.6, label=f"クラスタ{i}")
centers = kmeans.cluster_centers_
self.ax.scatter(centers[:, 0], centers[:, 1],
marker="*", s=200, color="white", zorder=5)
self.ax.set_title("K-Means クラスタリング (k=4)", color="#c9d1d9")
self.ax.tick_params(colors="#8b949e", labelsize=7)
for spine in self.ax.spines.values():
spine.set_edgecolor("#30363d")
self.ax.legend(facecolor="#161b22", labelcolor="#c9d1d9",
edgecolor="#30363d", fontsize=8)
self.fig.tight_layout()
self.mpl_canvas.draw()
inertia = kmeans.inertia_
self._set_report(
f"アルゴリズム: K-Means\nクラスタ数: 4\n"
f"慣性 (inertia): {inertia:.2f}\nサンプル数: {len(X)}")
self.status_var.set(f"完了: K-Means 慣性 {inertia:.2f}")
def _set_report(self, text):
self.report_text.configure(state=tk.NORMAL)
self.report_text.delete("1.0", tk.END)
self.report_text.insert("1.0", text)
self.report_text.configure(state=tk.DISABLED)
if __name__ == "__main__":
root = tk.Tk()
app = App079(root)
root.mainloop()
6. ステップバイステップガイド
このアプリをゼロから自分で作る手順を解説します。コードをコピーするだけでなく、実際に手順を追って自分で書いてみましょう。
-
1ファイルを作成する
新しいファイルを作成して app079.py と保存します。
-
2クラスの骨格を作る
App079クラスを定義し、__init__とmainloop()の最小構成を作ります。
-
3タイトルバーを作る
Frameを使ってカラーバー付きのタイトルエリアを作ります。
-
4入力フォームを実装する
LabelFrameとEntryウィジェットで入力エリアを作ります。
-
5処理ロジックを実装する
_execute()メソッドにメインロジックを実装します。
-
6結果表示を実装する
TextウィジェットかLabelに結果を表示する_show_result()を実装します。
-
7エラー処理を追加する
try-exceptとmessageboxでエラーハンドリングを追加します。
7. カスタマイズアイデア
基本機能を習得したら、以下のカスタマイズに挑戦してみましょう。
💡 ダークモードを追加する
bg色・fg色を辞書で管理し、ボタン1つでダークモード・ライトモードを切り替えられるようにしましょう。
💡 データの保存機能
処理結果をCSV・TXTファイルに保存する機能を追加しましょう。filedialog.asksaveasfilename()でファイル保存ダイアログが使えます。
💡 設定ダイアログ
フォントサイズや色などの設定をユーザーが変更できるオプションダイアログを追加しましょう。
8. よくある問題と解決法
❌ 日本語フォントが表示されない
原因：システムに日本語フォントが見つからない場合があります。
解決法：font引数を省略するかシステムに合ったフォントを指定してください。
❌ ライブラリのインポートエラー
原因：必要なライブラリがインストールされていません。
解決法：pip install コマンドで必要なライブラリをインストールしてください。 (pip install pandas matplotlib)
❌ ウィンドウサイズが合わない
原因：画面解像度や表示スケールによって異なる場合があります。
解決法：root.geometry()で適切なサイズに調整してください。
9. 練習問題
アプリの理解を深めるための練習問題です。
-
課題1：機能拡張
機械学習デモアプリに新しい機能を1つ追加してみましょう。
-
課題2：UIの改善
色・フォント・レイアウトを変更して、より使いやすいUIにカスタマイズしましょう。
-
課題3：保存機能の追加
処理結果をファイルに保存する機能を追加しましょう。