A Case Study on Analyzing the Similarity Between Two Similar Stocks: Building the CrC-LSTM Model

Nowadays, there is a huge innovation change in financial sector that requires complex digital technologies. For stock market investors, a good algorithm and model can output more accurate stock price for prediction of the future. This allows them to have a better portfolio management. Since the stock market price is a time-series data, most researchers tend to use algorithms which relates to the Long Short-Term Memory (LSTM) from Recurrent Neural Network. The forget gate in LSTM allows us to determine what kind of previous information can be stored and transferred. However, the original model only relies on its own past data. Thus, we start to implement a Cross-Correlated-LSTM that has a fourth gate named as the Link Gate. The aim of this link gate is to extract useful information from other similar time-series data. In this paper, we present the logic behind the CrC-LSTM via a case study by analyzing the similarity of the two candlestick graphs between the Apple Inc. and the Meta Platform Inc. combined with a dry run of the CrC-LSTM and obtained some results. The analysis and the prototype results have shown the potential of the CrC-LSTM in predicting the stock prices.