def load_data(): return getDataMatrix(TRAIN_FILES)
scores.append(score) #train_score = np.sum(np.abs(clf.predict(X_scaled) - y)) / len(y) train_score = np.sum(clf.predict(X_scaled) == y) / len(y) train_scores.append(train_score) lengths.append(len(y)) print("{:.4}\t\t{:.4}".format(train_score, score)) print( np.sum(np.array(scores) * np.array(lengths))/np.sum(lengths)) print(np.mean(scores)) """ if __name__ == "__main__": results = [] X, y = getDataMatrix(TRAIN_FILES) scaler = preprocessing.Scaler() scaler.fit(X) for i, f in enumerate(TRAIN_FILES): X, y = getDataMatrix(TRAIN_FILES[:i] + TRAIN_FILES[i+1:]) X_scaled = scaler.transform(X) clf = linear_model.LogisticRegression(C=0.04) clf.fit(X_scaled, y) s1 = test(f, use_classified_shot=True) s2 = test(f, use_classified_shot=False) #s3 = test(f, use_classified_shot=False, use_history=False) print(("{:.2%}\t\t"*2).format(s1, s2)) results.append((s1, s2)) r = np.array(results)