Building and Evaluating an LSTM Model for Time-Series Regression | S-Logix

List of Topics:

How to Build and Evaluate an LSTM Model for Regression Using Time-Series Tabular Data

LSTM Model for Regression

Condition for Building and Evaluating an LSTM Model for Time-Series Regression

Description:
This code preprocesses a household power consumption dataset by handling missing values and extracting date and time features. It then uses an LSTM model for regression, predicting the 'Global_active_power' based on other features. The model is evaluated using performance metrics like Mean Absolute Error (MAE), Mean Squared Error (MSE), and R-squared (R²).

Step-by-Step Process

Import Libraries:
Import essential libraries like numpy, pandas, tensorflow, sklearn, and matplotlib.
Load and Preprocess Data:
Load the dataset, handle missing values, and convert date and time columns.
Feature Extraction:
Extract relevant features such as year, month, day, hour, minute, second, etc.
Data Scaling:
Use StandardScaler to scale the data before feeding it into the LSTM model.
Build and Train Model:
Define and compile the LSTM model with two LSTM layers followed by a dense layer.
Evaluate Model:
Evaluate model performance using MAE, MSE, RMSE, and R² metrics.

Sample Source Code

# Import Necessary Libraries
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns
from sklearn.preprocessing import StandardScaler
from sklearn.model_selection import train_test_split

import warnings
warnings.filterwarnings("ignore")

from tensorflow.keras.layers import Input, LSTM, Dense
from tensorflow.keras.models import Model
from sklearn.metrics import mean_absolute_error, mean_squared_error, r2_score
import numpy as np

df = pd.read_csv("/path/to/dataset.csv", sep=';', header=0, low_memory=False)

# Check Nan values
print("Check Nan values\\n")
print(df.isna().sum())

# If Nan values present
df = df.dropna()

# Check Null Values
print("Check Null Values\\n")
print(df.isnull().sum())

# Convert Date and Time Columns
df['Date'] = pd.to_datetime(df['Date'])

df['Time'] = pd.to_datetime(df['Time'], format='%H:%M:%S').dt.time

# Continue as per original code...

Screenshots

List