|
|
 |
|
|
|
|
|
|
 |
|
|
|
|
|
TradingSolutions FAQ
When you see 
you can click on it to view a video for more information on the related topic.
- What is TradingSolutions?
- What are artificial intelligence, neural networks and genetic algorithms?
- Do I need the knowledge of a mathematician or scientist to use neural networks?
- Does TradingSolutions prohibit me from manipulating the neural network algorithms?
- How can I use TradingSolutions to determine the best time to buy and sell?
- What type of information can be fed into neural networks when searching for relationships in the data?
- What is the accuracy of neural networks for predicting future stock market movement?
- Why should I choose TradingSolutions over other neural network software packages?
- What is the required hardware to run neural networks in TradingSolutions?
- Why are you selling the software rather than making millions using it?
A. TradingSolutions is a software "toolbox" for building trading systems. It is not a trading system in its own right, it is a "toolbox" of traditional technical analysis, neural networks and artificial intelligence (AI) techniques you can combine to form intelligent trading models. The models can consist of indicators and rules like traders have used for years, artificial intelligence techniques, or hybrids of both. It will build models for stocks, futures, FOREX, commodities, indices and more. You can build models for exchanges all over the world including NYSE, AMEX, ASX and many more. To build models you just need to be able to obtain data for the instrument or exchange in which you are interested. Then the models you build will automatically back-test, and continue to give signals into the future as new data arrive.
Although we've made neural network modeling easier to use, it does not instantly build wonderful trading systems. However, we do provide a number of examples and other traditional systems designed by others. In addition, each new user of TradingSolutions will receive a unlimited access to our free Solution Service models so you will have working models to use while you build your own.
Q.You mentioned previously about artificial intelligence, neural networks and genetic algorithms. Could you please explain the differences between them and how each work?
A. Because there have been many significant advancements lately, artificial intelligence and neural networks are making quite a few headlines - and will continue to do so as they impact more aspects of our lives. While the math and science underlying each are fairly complicated, understanding the basics and applying them in user-friendly programs such as TradingSolutions, is relatively simple. Here's an overview of each:
Artificial Intelligence Neural Networks Sound complicated? What's important to understand is that a neural network allows your computer to sort through large amounts of data and recognize and match complicated, vague or incomplete patterns. For this reason, neural networks are well suited for projecting what will most likely happen next. TradingSolutions makes it easy to utilize this benefit of neural networks without requiring you to understand the details of the work being done behind the scenes.
Genetic Algorithms Genetic algorithms can be combined with neural networks to enhance their performance by taking some of the guesswork out of choosing the best neural network parameters and inputs. They can also be used to increase the opportunities of your trading signals.
Q. I see how neural networks can improve my trading results, but do I need the knowledge of a mathematician or scientist to figure out how to effectively apply neural networks?
A. Not at all! While the theory behind neural networks is rather complex, the interface in TradingSolutions makes it simple for anyone to apply them to their trading. In fact, after completing the step-by-step tutorial, you'll immediately be ready to use the prediction wizards to start making more intelligent trading decisions. But you don't need to take our word for it. You can download a free evaluation of TradingSolutions, complete the step-by-step tutorial, and experience the power and simplicity of neural networks risk-free.
Q. What if I am already familiar with neural networks? Does TradingSolutions prohibit me from manipulating the neural network algorithms?
A. Not at all! While TradingSolutions' prediction wizards make it easy to create predictions without having to know anything about neural networks, nothing prevents you from looking under the hood and fine-tuning the settings to your satisfaction. Advanced panels are available to allow you to directly change settings such as the number of processing elements and learning rates. You can even implement your own neural network designs using our NeuroSolutions and Custom Solution Wizard products that come bundled with the TradingSolutions Suite.
Q. Could you give me an example of how I can use TradingSolutions' neural networks to determine the best time to buy and sell my technology stocks?
A. Here are a couple of examples to illustrate how you could use neural networks:
TradingSolutions has a unique feature that allows you to analyze historical data to determine when the best buying and selling opportunities would have been. TradingSolutions calls this the Optimal Signal Another common way to use neural networks in financial markets is Predicting Prices
Q. What type of information can be fed into neural networks when searching for relationships in the data?
A. Whether you are predicting the future prices or modeling the Optimal Signal, TradingSolutions neural networks can identify their relationships to any date-based information. This includes basic trading information (e.g. high, low, open, close and volume), technical indicators
Q. Can you tell me what the potential accuracy of neural network systems is for predicting future stock market movement?
A. When testing on out-of-sample data (which means data that has not been previously used for modeling), the majority of neural network models that we trade have an accuracy rate of 60 to 85 percent *. Please see our sample models at (www.tradingsolutions.com/products/sample.html) to check out the performance of some of the models we have created using TradingSolutions.
Q. Why should I choose TradingSolutions over other neural network software packages?
The differences are immediately apparent the first time you run TradingSolutions - professionalism, power, and ease of use.
We are often asked how we compare with other trading software. Quite frankly, we don't have the time to keep up with what everyone else is doing, but we are confident enough to offer a free evaluation of TradingSolutions to prove how powerful it really is unlike other competitors. What we do know is that no other package makes it so easy to manage the analysis of all of the stocks you want to keep track of. Easy-to-understand wizards guide you through each task. Built-in charting and spreadsheet tools allow you to view the data without having to switch to other software. And, comprehensive helps and tutorials answer everything you want to know about the package when you need to know it.
Underneath the hood, TradingSolutions incorporates something you won't find in other packages -- time-lag recurrent neural networks. These networks were conceived as a result of years of research into the prediction of future values using historical time-based data. Furthermore, exclusive to TradingSolutions is the ability to model optimal trading signals as discussed in a previous answer. Other unique features include the ability to create committees of successful trading signals, the ability to work with the neural networks at any level of expertise, and an advanced correlation analysis wizard to assist you in selecting informative inputs for your predictions.
Q. Because neural networks are so advanced and powerful, do they require all the latest hardware to run them?
A. As with any software program, the more advanced your system, the faster it will run. But you don't need a high-end machine to put neural networks to good use. In fact, TradingSolutions can still operate on any Pentium or AMD processor, although we recommend a 2GHz or faster computer.
Q. If TradingSolutions is so powerful, why are you selling the software rather than making millions using it?
A. Actually, several members of our staff use TradingSolutions on a regular basis, and quite successfully. We in turn use our trading experiences to enhance the software, and this benefit gets passed along to you. Our specialty is creating high-quality, cutting-edge software. Furthermore, we have worked with several established financial institutions on a consulting or partnership basis in solving problems or developing custom solutions to their ideas. While, we feel that we can use this software to increase our own trading opportunities, we don't see a reason to keep others from sharing in the knowledge too.
* Hypothetical performance results have many inherent limitations. No representation is being made that any account will or is likely to achieve profits or losses similar to those shown. In fact, there are frequently sharp differences between hypothetical performance results and the actual results subsequently achieved by any particularly trading program.
One of the limitations of hypothetical performance results is that they are generally prepared with the benefit of hindsight. In addition, hypothetical trading does not involve financial risk. Variables such as the ability to adhere to a particular trading program in spite of trading losses as well as maintaining adequate liquidity are material points which can adversely affect actual real trading results.
Artificial intelligence is a term used to refer to a broad category of technologies that allow a computer to simulate learning or subjective reasoning. It includes neural networks, genetic algorithms, expert systems and other approaches that you may have heard of. Each of these technologies has different strengths and weaknesses that make them good at solving particular problems.
Neural networks offer some exciting prospects because they can actually learn from data to train themselves to perform more effectively. Neural networks try to model how the biological human brain works by simulating the basic functions of natural neurons. Artificial neural networks consist of layers of artificial neurons, which are then connected to each other with varying parameters that represent the strengths of these connections. The networks "learn" when they adjust these strengths to cause the overall network to improve its results.
Genetic algorithms are specialized computer programs that are able to efficiently solve a problem by intelligently searching through a large number of potential solutions for that problem. They are typically able to find good solutions to a problem much faster than traditional search methods. The idea behind genetic algorithms is to simulate the way nature uses evolution by applying the "Survival of the Fittest" principle to small sets of its potential solutions. The good solutions reproduce to form new and hopefully better solutions to the problem, while the bad solutions are removed.
. The neural networks can then be used to analyze this buy and sell information for relationships with other factors, such as the prices for preceding days. Given these relationships, the neural network can help to determine what position you should currently be in for optimal opportunities. The idea is that if you can model where you should have been in the market throughout history, you can predict where you should be in the market right now.
, such as the closing price of Microsoft stock, and base a buy and sell strategy on that predicted price. First, you download a set of data for Microsoft, such as a historical list of prices it has traded at. Then, you allow the neural network to analyze the data for relationships between the closing price and other factors, such as the prices for the preceding days. Once this is done the neural network can be used to predict future closing prices, allowing you to identify buying opportunities.
(e.g. moving average, stochastic oscillator, relative strength index, etc.), and broad market conditions (e.g., the Dow Jones Industrial Average, TRIN, etc.). Whatever you think is relevant can be fed into the neural networks to improve the predictions.
 |
 |