PREDICTING STOCK BEHAVIOUR USING VAR AND NNAR MODELS: ACOMPARATIVE ANALYSIS

Abstract

In the modern day predicting stock market behavior has become important because data-driven decisions help traders, analysts, and policymakers a lot with their day-to-day life which impact their portfolios. This research aims to use Vector Auto-regressive (VAR) and Neural Network Auto-regressive (NNAR) models to analyze stock market benchmarks that affect the sentiment of traders and the market. By examining data and interrelated financial variables the project aims to provide forecasts that reveal trends and risks involved in investing in a particular stock. Utilization of VAR and NNAR models will help our understanding of the future ahead of us by providing insights for decision-making. This study not only contributes to the field of analysis but also has practical applications concerning investment strategies, risk management, and policy development. The VAR model is a multivariate time-series model that captures the relationships by relating current observations of a variable with the past observations of itself and previous values of other variables in the system over the lagged values. The NNAR model is a feed-forward neural network with lagged inputs and a single hidden layer with non-linear functions it is said to be one of the best tool for analyzing time-series data.