The Application And Study Of Plan Recognition In Stock Trends Analysis

Since the stock market has been born, people want to analysis and forecast the futuretrend of the stock and obtain interests, according to various factors and signals come fromthe stock market. From the earliest irrational “skirt theory” to scientific theory“Fundamentals analysis” and“Technical analysis” methods, which use respectively statisticsand graphics. These theories get certain effect in different periods. In the new era, computertechnology has been widely applied to various fields with the development of computertechnology, and show good results. So people hope that this new technology can be appliedto the analysis of the stock market, so as to help us obtain the benefit and reduce risk in thestock market.For this, many scholars have made a lot of research work in this field. Such as, neuralnetworks, data mining methods and other artificial intelligence methods are applied topredict the stock market, and achieved certain results. However, most of these methodsspeculate on the future price changes, according to historical experience which gets throughresearching the historical data of the stock market. But there is a lot of cheats and game inthe stock market, so it is difficult to get a reasonable inference simply according to thehistorical dataPlan recognition method will be applied to the stock market in this article. The problemof plan recognition is a cross subject from psychology and artificial intelligence. It relates toknowledge representation, knowledge reasoning, no monotonic logic, the situation calculus,human interaction, and knowledge mining and so on. There are many methods in planrecognition system. In those methods, plan recognition based on abstract strategy is verysuitable for stock market. The method is mainly used in the noise and uncertainty domain. Itwill establish the multi-layer reasoning. First，Marco model is used as a probabilisticinference in the probabilistic inference layer, and each different investment action willproduce different effect in the stock market investment. So the utility theory is introducedinto the second layer, and gets a more reasonable decision from using the method ofdecision theory. Last layer, we will reduce the complexity of computation according to thepolicy iteration layer. The last of this paper, the validity of the strategic planning method isverified, and have achieved good results. We prospect the future of plan recognition inanalysis of stock market through the summary and analysis of this paper.