Time Series Forecasting and Prediction of Walmart Data using Hybrid Machine Learning Techniques

Abstract

Time series forecasting gained attention being a popular technique widely used in different industries of finance, production, business, supply chain management, production, and inventory planning. When making predictions about a particular problem, it often involves analyzing data that varies over time, and this requires time series forecasting. Different machine learning techniques such as neural networks, support vector machines(SVM), random forests, and regression are used for making predictions. Essentially, forecasting involves building models using past data, and then using these models to make predictions about what may happen in the future. In this paper we have proposed a hybrid model by combining Convolutional Neural Network (CNN) and Artificial Neural Network (ANN). Then compared performance of proposed model with existing model such as Long Short-Term Memory (LSTM), Multi-Layer Perceptron (MLP), ANN and CNN. Our simulation result shows that proposed model outperforms existing technique using parameters Root Mean Squared Error (RMSE) and Mean Absolute Error (MAE).