Machine Learning Predictive Model For Time Series Data and Time Series Forecasting Problem

The problem of Time Series data requires a special approach for data analysis, and pre-processing. Based on dataset exploration, it should be chosen Machine Learning Models which could resolve Time Series Forecasting problem. After analyzing different approaches and algorithms, it was decided to use Long Short-Term Memory models with different parameters that have the Recurrent Neural Network architecture with purpose to train and test dataset and to compare test results and predictions. Received testing results satisfy expectation and can be applied for resolving Time Series prediction problems.