krish517/Temperature-prediction-using-different-machine-learning-models: Model Validation for more than a year of data

Release Notes – Initial Repository Version Overview This release provides a complete implementation of the machine learning framework developed for location-specific time-series prediction. The repository contains all source code, datasets, and model configurations necessary to reproduce the experiments. The workflow includes data preprocessing, model training, testing, and performance evaluation using long-term hourly environmental data. Repository Structure Code: Includes scripts for data preprocessing, feature engineering, model training, hyperparameter tuning, evaluation, and visualization. Data: Contains hourly datasets systematically organized by geographic location, covering the years 2002–2023. Models: Provides implementations and stored weights for four predictive models: Extreme Gradient Boosting (XGB) Random Forest (RF) K-Nearest Neighbors (KNN) Long Short-Term Memory network (LSTM) Data Description Training Data: Hourly observations collected from 2002 to 2022. Testing Data: Hourly observations for April 2022 to May 2023. All datasets have been preprocessed, normalized, and structured to ensure compatibility across all model types. Model Details XGB, RF, and KNN: Traditional machine learning models optimized using location-dependent features and evaluated for robustness across multiple regions. LSTM: A deep learning model designed to capture temporal dependencies and nonlinear patterns inherent in long-term hourly data. Key Features Comprehensive handling of location-wise datasets. Standardized evaluation procedures, including performance metrics and visualization outputs. Fully reproducible experimental pipeline enabling systematic comparison of multiple models. Modular code structure supporting easy extension and adaptation for future research. Version Notes This release represents the first stable version of the repository. It includes all essential components required for conducting time-series prediction experiments using machine learning and deep learning models. The repository is intended to support researchers and practitioners working in the areas of time-series forecasting, environmental modeling, and geospatial data analysis.