Stock trading and prediction using neural network

Stock market prediction has been an area of great interest to financial researchers and practitioners. Various prediction techniques have been applied in time series forecasting. Recently, artificial NNs (NNs) have been popularly applied in these area due to its ability to find patterns and irregularities as well as detecting multi-dimensional non-linear connections in data. Many researches have been conducted in the past to investigate its performance as the stock market prediction model, and encouraging results are found. In this project, a two phases NN modeling method is proposed, developed and evaluated. The modeling method consists of the first building of preliminary prediction model for technical indicators parameters optimization, and the second building of final prediction model using the optimized technical indicators. Genetic Algorithm (GA) is used to apply in the optimization and a stop loss strategy is also designed and further integrated to the final model and effectively improves the profitability of the model. Technical indicators are use to interpret and convert raw stock prices and volume into discrete value which better representing the market condition. The NN model takes 11 inputs generated from the technical indicators and produces its output based on the recognition of input pattern. Finally, buy/hold/sell signal are generated based on the NN output value. The Proposed method is compared with the IPLR model developed by Chang et al. [3] and also further evaluated using seven major Asia/Pacific stock indexes. The experimental results show that the proposed method is able to generate promising rate of return.Bachelor of Engineerin