FORECASTING OF TIME SERIES DATA USING PROPHET: A GRANULARITY BASED APPROACH
Manjula. K. A, P Karthikeyan. · INTERNATIONAL JOURNAL OF ADVANCED RESEARCH IN ENGINEERING & TECHNOLOGY · 2020
Time series forecasting is used to predict future values based on previously observed data points collected over time at regular intervals.This study uses the Prophet model to examine how data granularity affects time series forecasting performance.We trained models for prediction using historical stock price data from 2008 to 2018 and predicted market movements for 2019 with a focus on nine significant Indian IT businesses.Training was performed using daily, weekly, and monthly data of ten years and prediction was done for a future period of one year.The Forecast accuracy was examined using the Root Mean Square Error (RMSE) metric.The findings show that more accurate forecasts are produced by finer granularities (daily data), with RMSE rising with coarser data.This study provides insights for analysts and investors and highlights the significance of data granularity in financial forecasting.