A new method for predicting stock market crashes using classification and artificial neural networks. (15th July 2020)

• Record Type: Journal Article
• Authors: Tabar, Saeed; Sharma, Sushil; Volkman, David
• Is Part Of: International journal of business and data analytics. Volume 1:Number 3(2019)
• Page Start: 203
• Page End: 217
• Publication Date: 2020-07-15
• Subjects: artificial neural networks -- ANNs -- classification -- stock market -- market prediction -- market crash; Commercial statistics -- Data processing -- Periodicals; Industrial management -- Mathematical models -- Periodicals; Business -- Mathematical models -- Periodicals; Management -- Statistical methods -- Periodicals; Business -- Research -- Periodicals; 658.403
• Languages: English
• ISSNs: 2515-9100

Abstract

The stock market prediction is an interesting topic, especially for traders and investors. One important aspect of predicting the stock market is identifying price patterns which may result in a market crash. With the advancement of computer technology, particularly in the area of artificial intelligence, a large number of new models have been proposed. The proposed method in this article is based on identifying the normal behaviour of a crowd in the stock market using exponential moving average and then classifying the price fluctuations into three categories BUY, SELL, and STOP. An artificial neural network (ANN) with five input neurons, ten hidden neurons, and three output neurons is then used to learn from the price fluctuations and predict one day ahead. The final results show that the algorithm is capable of identifying the market crashes in advance by issuing STOP labels.