Time Series Forecasting — ML & DL End-to-End Project

📌 Overview:

A practical time series forecasting project applying both traditional Machine Learning techniques and advanced Deep Learning architectures. The repository includes full data preprocessing, visualization, and prediction workflows on real-world time series datasets: avocado prices and vehicle miles traveled.

📂 Project Structure:

Time_Series_Forcasting(TSF)ML.ipynb

👉 Forecasting avocado prices using:

Linear Regression

Random Forest Regressor

Prophet (Facebook’s time series model)

📌 Includes:

Data loading & sorting by date.

Data exploration (info(), describe()).

Handling time series ordering.

Applying and evaluating ML models.

Model comparison using MAE / MSE.

Time_Series_Forcasting_(TSF)_DL.ipynb

👉 Forecasting vehicle miles traveled using:

LSTM Neural Network (Keras Sequential model)

📌 Includes:

Data exploration & visualization (plot()).

Setting time-based index.

Feature scaling using MinMaxScaler.

Data reshaping for LSTM input.

Building and training an LSTM model.

Forecasting and plotting predictions.

Evaluating performance via MSE / MAE.

📊 Tools & Libraries: Python

Pandas, Numpy

Matplotlib, Seaborn

Scikit-learn

Keras (TensorFlow backend)

Prophet

📈 Key Highlights:

Real-world time series datasets.

End-to-end workflow from data preprocessing to model training and prediction.

Clear visualizations to interpret trends and model performance.

Hybrid use of classical ML regressors and sequence-based DL models.

Hands-on implementation of Prophet and LSTM for time series.