MLPRegressor Python 

การทำนายข้อมูลด้วย MLPRegressor (Train)

..

👉..

1. import pandas as pd

2. import numpy as np

3. import matplotlib.pyplot as plt

4. import pickle

5. import joblib

6. 
   

7. from sklearn.model_selection import train_test_split, learning_curve

8. from sklearn.preprocessing import StandardScaler

9. from sklearn.neural_network import MLPRegressor

10. from sklearn.metrics import mean_squared_error, r2_score

11. 
    

12. # 1. Load the data from a CSV file

13. data = pd.read_csv(r"C:\Users\user\Desktop\motor_log.csv")  # Replace with your CSV file path

14. 
    

15. # Suppose the CSV has features 'current', 'volt', 'vibration' and the target 'perform'

16. X = data[['current', 'volt', 'vibration']].values

17. y = data['perform'].values

18. 
    

19. # 2. Split the data into training and testing sets

20. X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)

21. 
    

22. # 3. Scale the features (important for neural networks)

23. scaler = StandardScaler()

24. X_train_scaled = scaler.fit_transform(X_train)

25. X_test_scaled = scaler.transform(X_test)

26. 
    

27. # 4. Create and train the nonlinear regression neural network using MLPRegressor

28. # Define a network with two hidden layers: first with 100 nodes, second with 50 nodes

29. mlp = MLPRegressor(hidden_layer_sizes=(100, 50),

30.                    activation='relu',

31.                    solver='adam',

32.                    max_iter=1000,

33.                    random_state=42)

34. 
    

35. for epoch in range(1000):  # Manually iterate over epochs

36.     mlp.partial_fit(X_train_scaled, y_train)

37.     print(f"Epoch {epoch+1}, Loss: {mlp.loss_:.4f}")

38. 
    

39. # Fit the model on the training data

40. mlp.fit(X_train_scaled, y_train)

41. 
    

42. # 5. Plot the learning curve

43. # The learning_curve function returns training sizes, training scores, and cross-validation scores.

44. train_sizes, train_scores, test_scores = learning_curve(

45.     mlp, 

46.     X_train_scaled, 

47.     y_train, 

48.     cv=2,  # Changed from 5 to 2 folds to avoid the error

49.     scoring="neg_mean_squared_error",  

50.     train_sizes=np.linspace(0.1, 1.0, 10)

51. )

52. 
    

53. # Convert negative MSE to positive values by negating the scores for plot data

54. train_errors = -np.mean(train_scores, axis=1)

55. test_errors = -np.mean(test_scores, axis=1)

56. 
    

57. # 6. Plot the learning curve

58. plt.figure(figsize=(10, 6))

59. plt.plot(train_sizes, train_errors, 'o-', color="r", label="Training Error")

60. plt.plot(train_sizes, test_errors, 'o-', color="g", label="Cross-Validation Error")

61. plt.xlabel("Training Set Size")

62. plt.ylabel("Mean Squared Error")

63. plt.title("Learning Curve for MLPRegressor")

64. plt.legend(loc="best")

65. plt.grid(True)

66. plt.show()

67. 
    

68. # 7. Save the model and scaler to disk

69. with open(r"C:\Users\user\Desktop\mlp_model.pkl", "wb") as file:

70.     pickle.dump(mlp, file)

71. joblib.dump(scaler, r"C:\Users\user\Desktop\scaler.pkl")

72. print("Model and scaler saved successfully!")

เมื่อรันโปรแกรมแล้วผลลัพท์ที่ได้