Introduction ___________________________________________________________ 2 My Data: __________________________________________________________________ 2 Price: __________________________________________________________________________ 2 P/E Ratio:_______________________________________________________________________ 2 Volume: ________________________________________________________________________ 3 Williams %R:____________________________________________________________________ 3 My stocks _________________________________________________________________ 4 General Electric: _________________________________________________________________ 4 3M:____________________________________________________________________________ 5 Wal-Mart:_______________________________________________________________________ 6 Microsoft:_______________________________________________________________________ 6 Neural Network Program: ___________________________________________________ 7 Analysis and Preparation of Data__________________________________________ 7 Data Sets Chosen:___________________________________________________________ 9 Laying out the data:_________________________________________________________ 9 Experimenting ________________________________________________________ 10 Initial Results: ____________________________________________________________ 10 Refining: _________________________________________________________________ 10 Final Settings:_____________________________________________________________ 11 Final Results _________________________________________________________ 12 Results: __________________________________________________________________ 12 Reversed Data; ____________________________________________________________ 12 Training with GE: _________________________________________________________ 13 Baseline Test: _____________________________________________________________ 13 Explanation of Results__________________________________________________ 14 3M Results:_______________________________________________________________ 14 Wal-Mart Results: _________________________________________________________ 15 GE Resultes: ______________________________________________________________ 15 Conclusions:__________________________________________________________ 15 Buy of Sell? __________________________________________________________ 16Introduction For my project I wanted to use a back propagation neural network to decide if you should buy or sell a stock on a one year outlook. I’m something of a stock addict. I’ve been playing the stock market since I was in eighth grade. During the early years you could probably classify me as a near day trader (school required me to take a few days off now and then). As the market went south I managed to thrive, but I came to realize I’d have gray hair by the time I was 30 if I had any hair left if I continued day trading. Since that point I’ve been trying to make myself into a long term investor. When I thought about predicting the stock market for this project I realized that I could use it as an opportunity to try to get myself into the right long term mindset. I choose a year time frame; which is about as long term as I can think about right now, when you are use to trading times measured in days if not hours a year is a very long long time. I also set a goal that I wanted, using results from the last twenty years to be able to predict how the market performed during both the bull market of the late nineties and the collapse around 2000. Based on this I choose to use the years 1996 – 2003 as my testing zone. I was also interested in finding out how stocks in very different sectors compared to one another in terms of patterns the network could detect. I’ve always had a theory that regardless of what sector you are in on average stocks performs similarly in similar situations. My final goal for this project was to prove that you don’t need to use a hundred different numbers to get decent results. I’ve often found that there are two kinds of investors, those who get bogged down by the thousands of different stats you can get about a stock and those who just ignore all the numbers. Neither approach seems right to me. New traders are often highly intimidated by all the data they are presented with. When I first started to trade I use to be a numbers man. I would keep track of dozens of indicators for each stock I followed. Over the years though I’ve come to realize that it is much more important to know what a few select numbers are telling you then knowing what current stats are for all of them. So for this project I decided I would base all of my calculations off of just four pieces of data. All though this sounds extreme; I’ve been very successful in the stock market these last few years by following only a few more indicators. I suspected that a neural network would need even fewer data points to find a pattern. Based on my results I would have to say my hypothesis was justified. My Data: Price: This is probably the most obvious choice for stocks. The price you pay for a stock and the price you sell a stock at determine how well you do in the stock market. P/E Ratio: This is my favorite indicator. P/E stands for price to earnings. The Price to Earnings ratio is calculated as the current market price of a company's common stock divided by that company's earnings per share in the previous 12-month period.The PE ratio allows for the simplest comparison between different shares, as companies within a particular industry generally fall within a certain PE range. Comparisons between companies in different industries, however, are generally not appropriate using the Price to Earnings ratio. I learned the hard way during the dot bust that a PE ratio of a 100 really isn’t a deal. Through the years I’ve found that PE often is one of the best indications of whether a stock is overbought of oversold for the long run. When you see rapid changes in PE Ratio I’ve found that it is normally a good time to start looking into either buying or selling depending on the situation. Volume: Volume plainly put is the number of shares bought and sold for a given period. A large percentage price increase accompanied by a higher than average volume is a strong indicator of future price movements. A large percentage price movement accompanied by lower than average volume is a very weak indicator of
View Full Document
Unlocking...