Forecasting of Major World Stock Exchanges Using Rule-Based Forward and Backward Chaining Expert Systems

Abstract

Nowadays, share price forecasting has been considered as vital financial problem and received a lot of attention from financial analyst, researchers, brokers, stock users, etc. For the last couple of years, the research in stock market field has grown to a large extent. Therefore, artificial intelligence (AI) is a well-known and more popular area to this field. In AI, specifically expert system (ES) is a thought-provoking technique which has mystery to mimic human abilities to solve particular problems. This study mainly employs forward chaining- and backward changing-based expert system inference approaches to forecast the behavior of major stock exchanges like India, China, the USA, Japan, etc. However, the various financial indicators like inflation rate, foreign direct investment (FDI), gross domestic product (GDP), etc. have been considered to build the expert knowledge base. Moreover, the common LISP 3.0-based editor is used for expert system testing. Finally, experimental results show that backward chaining performs well when a number of rules and facts are large as compared to forward chaining approach.