This action might not be possible to undo. Are you sure you want to continue? Rule Based Approach for Stock Selection: Artificial Intelligence AI is one of most emerging field in the research field and has given boost to the most growing technology known as Expert System ES.
An ES is a computer based program that is bu ilt to solve the problems in specialized domain. There are various expert systems in diversified areas like medical, engineering, agriculture, science etc. This paper mainly highlights the need of expert system in stock market as well as rule based expert system approach for particular stock selection. However it is not surprising that lots of work has to be done on accurately predicting the stock market trends.
Over the last six decades various methods have been proposed for stock price prediction but no method or combination of methods have been succeeded to beat the volatility of the market .
In traditional stock market prediction system investors or brokers analyze the stock market behavior i. Now traditional approach is replaced by expert system approach which will work intelligently and efficiently for stock market and will guide us to know time to time market conditions . Architecture of Rule Based Expert System.
An expert system is a computer program that attempts to mimic human capability by the system capability to learn, to advice, to teach and performed intelligent tasks . A rule based expert system is an expert system where knowledge is encoded by series of rules. The major components of rule based expert system are : The reasoning mechanism where expert system reached out a certain decision and solve a problem.
The user interface provides the communication between user and expert system. It provides an useful external environment i.
It usually includes knowledge base editors, debugging aids i. In these days stock price prediction is an important issue. To making the right prediction useful knowledge is stored in knowledge base and knowledge is collected from various sources like stock websites, newspaper, financial institutes, books, magazines, T V channels etc. After collecting the knowledge the organization and representation of this one is also important concern.
For this various encoding schemes and knowledge representation techniques used. Sometimes also provide explanation or justification to the user i. Architecture of Rule Based Expert System . Their expert system successfully predicted all the events of the business. It would be very helpful to draw conclusions based on events and might be look forward or backward. Their system also has a capability to change data over time.
Boer and Livnat  have proposed an expert system for financial applications. Their expert systems are most widely used to train managers, financial experts and other industrial analysts.
The knowledge based to be used in such a system very effective and updated time to time. Smith and McDuffie  has used various quantative ratios like profitability, total profit, long term and short term debt etc. Finally their system showed that better decision making before investment. The proposed model used indirect approach fuzzy modeling.
They also proposed clustering approach for automatic rules extraction. The input variables risk, return, dividend etc. The proposed model had succeeded to particular portfolio selection for stock users and most promising results are generated in real time trading environments.
The model performance compared with other models like Sugeno-Yasukava model, multiple regression models, and feed-forward neural network model. The comparison demonstrated that proposed model had great performances In terms of minimize error rate, robustness and flexibility. Merloti  has proposed fuzzy. Knowledge Engineer or Domain Expert. His expert system was used to show how many stocks to buy or sell based on input values like price and MAD indicator. He also put various research questions i.
They applied the metho d on IBM stock and compared the results with previous methods. Abdalla  has propos ed various artificial intelligence approaches for analyze the modern financial time series. They also used various tools for analyzing the stock market variability. Finally they concluded that artificial intelligence approaches with combination of technical analysis could lead to significant performance.
Hamid and Vida  have proposed hybrid intelligent system for g as price forecasti ng. This study inspired by rule based expert system applications in stock market field. Stock market stores large amount of data. So it would be very effective to represent knowledge in terms of set of rules IF- THEN rather than declarative form or some in a static way. It would also help to reach at certain conclusions that are derived under given conditions or in diversified situations.
A rule based system comprise of  : Inference mechanism which decides which rule is to be applied based on the set of available facts and also shows the action against the selective rule.
Forward chaining is the data-driven reasoning. The reasoning process starts from the initial facts or known data and proceeds forward with that data.
Each time only top most rules are executed when fired and the rule adds a new fact in the database. The most important thing the rule can be executed only once.
The match —fire procedure stops when no further rules can be fired  . The working p rocedure of forward chaining approach is as follows: In first step suppose database contain some initial facts like A, B, C, D and E and knowledge base contain only five rules: In the second step only two rules i.
So Y is added in the database and this will in turn to execute Rule 1 i. Stock market contain so much variability in nature i. Sign up to vote on this title.
Read Free for 30 Days Cancel anytime. Read Free for 30 Days. Close Dialog Are you sure? Also remove everything in this list from your library. Are you sure you want to delete this list? Remove them from Saved? An Expert System will be available on Restart preview.More...