| Summary: | 碩士 === 國立高雄海洋科技大學 === 電訊工程研究所 === 102 === Stock investment is an indispensable modern financial tool. The main subject of this thesis is to predict the S&P 500 Index and Taiwan's financial stock index for the main purpose, but the share price volatility has always been a critical issue for investors, There are many factors that can affect the stock price, for example: political issues, financial crisis and so on. In order to provide the investors a tool to build a neural network to predict stock prices, and to reduce investment risk, many researchers work this important issue.
In order to effectively predict the S&P500 Index, we pre-processed the data first, excluding the higher-end and lower-end raw data. Then we used the Back - propagation Neural Network (BPNN) model to predict the highest-price, lowest-price, and the daily closing-price of S&P 500 Index. The raw data and pre-processed data were compared to find advantages and disadvantages of each methods. The data used in this research were collected from 3rd. January. 2000 to 2nd. November. 2012, 3056 data points in total. In this study, input variables include two-before items and one-before items’ opening-prices, highest-prices, lowest-prices and closing-prices. The data from 2000 to 2007 were used for training the algorithn; while those from 2008 to 2012 were used for testing.
For the sake of the large-number lay inrestors in Taiwan's stock market, this study want to discuss the prediction of Taiwan's financial stock Index. The data used in this research were collected from 31st. July. 2013 to 24th. February. 2014, 140 data in total, where the first 80 data are training samples, the following 60 data for testing sample.
This research first used neural networks in Support Vector Machine(SVM) to classify information, and took the points on the boundary as in the number of hidden layer matrix Radial Basis Function Network (RBFN). The input data variables include two-before items and one-before items’ opening-prices, highest-prices, lowest-prices and closing-prices to predict the highest-prices, lowest-prices, and the closing-prices for Taiwan Financial Index.
|
|---|
