epochs=EPOCHS,
validation_data=(X_valid, y_valid),
callbacks=[earlystop],
verbose=1)
# Test the model
X_test = test_inputs['X']
y1_test = test_inputs['target_load']
y1_preds = model.predict(X_test)
y1_test = y_scaler.inverse_transform(y1_test)
y1_preds = y_scaler.inverse_transform(y1_preds)
y1_test, y1_preds = flatten_test_predict(y1_test, y1_preds)
mse = mean_squared_error(y1_test, y1_preds)
rmse_predict = RMSE(mse)
evs = explained_variance_score(y1_test, y1_preds)
mae = mean_absolute_error(y1_test, y1_preds)
mse = mean_squared_error(y1_test, y1_preds)
meae = median_absolute_error(y1_test, y1_preds)
r_square = r2_score(y1_test, y1_preds)
mape_v = mape(y1_preds.reshape(-1, 1), y1_test.reshape(-1, 1))
# print("mse:", mse, 'rmse_predict:', rmse_predict, "mae:", mae, "mape:", mape_v, "r2:", r_square,
# "meae:", meae, "evs:", evs)
print('rmse_predict:', rmse_predict, "evs:", evs, "mae:", mae, "mse:", mse,
"meae:", meae, "r2:", r_square, "mape", mape_v)
predictions = model.predict(X_test)
# compare predictions with actual ones
eval_df = pd.DataFrame(predictions,
columns=['t+' + str(t) for t in range(1, HORIZON + 1)])
eval_df['timestamp'] = test_shifted.index
eval_df = pd.melt(eval_df,
id_vars='timestamp',
value_name='prediction',
var_name='h')
eval_df['actual'] = np.transpose(y_test).ravel()
eval_df[['prediction', 'actual'
]] = scaler.inverse_transform(eval_df[['prediction', 'actual']])
eval_df.head()
# Compute the mean absolute percentage error over all predictions
mape(eval_df['prediction'], eval_df['actual'])
# plot the result
eval_df[eval_df.timestamp < '2014-11-08'].plot(x='timestamp',
y=['prediction', 'actual'],
style=['r', 'b'],
figsize=(15, 8))
plt.xlabel('timestamp', fontsize=12)
plt.ylabel('load', fontsize=12)
plt.show()
## clean up the model
for m in glob('model_*.h5'):
os.remove(m)
print("done")
X_train = X_train.sort_index()
y_train = y_train.sort_index()
regressor = LinearRegression()
regressor.fit(X_train, y_train) #training the algorithm
y_pred = regressor.predict(X_test)
# print('Mean Absolute Error:', metrics.mean_absolute_error(y_test, y_pred))
# print('Mean Squared Error:', metrics.mean_squared_error(y_test, y_pred))
# print('Root Mean Squared Error:', np.sqrt(metrics.mean_squared_error(y_test, y_pred)))
# print("R2:", metrics.r2_score(y_test, y_pred))
predicted_Y_entire = regressor.predict(X)
print('Mean Absolute Error:', metrics.mean_absolute_error(y_test, y_pred))
print('Mean Squared Error:', metrics.mean_squared_error(y_test, y_pred))
print('Root Mean Squared Error:',
np.sqrt(metrics.mean_squared_error(y_test, y_pred)))
mape_v = mape(y_pred.reshape(-1, 1), y_test.values.reshape(-1, 1))
print('mape:', mape_v)
r2 = metrics.r2_score(y_test, y_pred)
print("R2:", r2)
store_predict_points(
Y_entire, predicted_Y_entire,
'output/test_linear_without_prediction_r2_' + str(r2) + '.csv')
# df = pd.DataFrame({'Actual': y_test, 'Predicted': y_pred})
# df.plot(kind='bar',figsize=(10, 8))
# plt.grid(which='major', linestyle='-', linewidth='0.5', color='green')
# plt.grid(which='minor', linestyle=':', linewidth='0.5', color='black')
# plt.show()
if index % 10 == 0:
print("rebuilding the model", index)
model = ARIMA(history, order=(time_step_lag, 1, 0))
model_fit = model.fit(disp=0)
output = model_fit.forecast()
yhat = output[0]
predictions.append(yhat)
obs = test[t]
history.append(obs)
print('predicted=%f, expected=%f' % (yhat, obs))
error = mean_squared_error(test, predictions)
print('Test MSE: %.3f' % error)
# plot
# plt.plot(test)
# plt.plot(predictions, color='red')
# plt.show()
mse = mean_squared_error(test, predictions)
rmse_predict = RMSE(mse)
evs = explained_variance_score(test, predictions)
mae = mean_absolute_error(test, predictions)
meae = median_absolute_error(test, predictions)
predictions = np.array(predictions)
r_square = r2_score(test, predictions)
mape_v = mape(predictions.reshape(-1, 1), test.reshape(-1, 1))
print("mse:", mse, 'rmse_predict:', rmse_predict, "mae:", mae, "mape:",
mape_v, "r2:", r_square, "meae:", meae, "evs:", evs)
look_back_dt = dt.datetime.strptime(
test_start_dt, '%Y-%m-%d %H:%M:%S') - dt.timedelta(hours=T - 1)
test = energy.copy()[test_start_dt:][['load', 'temp']]
test[['load', 'temp']] = X_scaler.transform(test)
test_inputs = TimeSeriesTensor(test, 'load', HORIZON, tensor_structure)
predictions = model.predict(test_inputs['X'])
eval_df = create_evaluation_df(predictions, test_inputs, HORIZON, y_scaler)
eval_df.head()
eval_df['APE'] = (eval_df['prediction'] -
eval_df['actual']).abs() / eval_df['actual']
eval_df.groupby('h')['APE'].mean()
print("MAPE:", mape(eval_df['prediction'], eval_df['actual']))
plot_df = eval_df[(eval_df.timestamp < '2014-11-08')
& (eval_df.h == 't+1')][['timestamp', 'actual']]
for t in range(1, HORIZON + 1):
plot_df['t+' + str(t)] = eval_df[(eval_df.timestamp < '2014-11-08')
& (eval_df.h == 't+' +
str(t))]['prediction'].values
fig = plt.figure(figsize=(15, 8))
ax = plt.plot(plot_df['timestamp'],
plot_df['actual'],
color='red',
linewidth=4.0)
ax = fig.add_subplot(111)
ax.plot(plot_df['timestamp'],
