Time Series Forecasting Uisng LSTM

Introduction

I have a data for past 10 years of 244 companies, data consist of ROE, ROA, REVENUE and w.r.t per year and same as BASIC and DILUTED EPS but both EPS are from 2015 to 2019. As we can see the nature of data is similar to time series forecasting for this purpose, I will use LSTM models for forecasting ROE, ROA, REVENUE and BASIC, DILUTED EPS. After forecasting I am classifying the companies in 3 categories Good, Worst and Neutral for example if a company ROE increase by 5% then its current ROE then it will be considered as Good if it drops then 5% then it will consider as Worst otherwise it lies in the neutral category. For each company I made two models in one model I use univariate approach and in other multivariate. I train the model using 2010 to 2019 data and the predict 2019 to 2025.

Click here to see Complete Report