Foreign Rate Exchange Prediction Using Neural Network and Sentiment Analysis

Foreign currency exchange plays a vital role for currency trading in the financial market. To manage large volume of transactions in modern world, it requires support from the computer algorithms. There could be potential problems like trading without a plan, having unrealistic expectation, failing to adapt to the market and many more. This paper examines on foreign exchange market prediction using neural network and sentiment analysis. There are various techniques and algorithms for prediction but different algorithms have different accuracy. Among them, one of the best and accurate method is Artificial Neural Network (ANN). Neural network parameters consist of number of neurons, use of bias neurons, number of hidden layers, activation functions and training methods. Root Mean Squared Error (RMSE) was found to be 0.0034 with 6 hidden nodes using ANN. As the price movement is also directly proportional to market sentiment, we applied sentiment analysis using combination of Naïve Bayes and lexicon based algorithm to analyze the opinion of different traders and predict the overall sentiment. Sentiments are taken from tweets and were classified as positive or negative. In sentiment analysis, accuracy was found to be 90.625%.