ium_495719/evaluate.py
2024-05-14 16:21:16 +02:00

39 lines
1.1 KiB
Python

import pandas as pd
import numpy as np
import sys
import os
from sklearn.metrics import mean_squared_error, mean_absolute_error, r2_score
from keras.models import load_model
from helper import prepare_tensors
build_number = int(sys.argv[1])
hp_test = pd.read_csv('hp_test.csv')
X_test, Y_test = prepare_tensors(hp_test)
model = load_model('hp_model.h5')
test_predictions = model.predict(X_test)
predictions_df = pd.DataFrame(test_predictions, columns=["Predicted_Price"])
predictions_df.to_csv('hp_test_predictions.csv', index=False)
rmse = np.sqrt(mean_squared_error(Y_test, test_predictions))
mae = mean_absolute_error(Y_test, test_predictions)
r2 = r2_score(Y_test, test_predictions)
metrics_df = pd.DataFrame({
'Build_Number': [build_number],
'RMSE': [rmse],
'MAE': [mae],
'R2': [r2]
})
metrics_file = 'hp_test_metrics.csv'
if os.path.isfile(metrics_file):
existing_metrics_df = pd.read_csv(metrics_file)
updated_metrics_df = pd.concat([existing_metrics_df, metrics_df], ignore_index=True)
else:
updated_metrics_df = metrics_df
updated_metrics_df.to_csv(metrics_file, index=False)