Hidden Variable Relation Discovery Algorithm Based On Parameter Space And Its Application In Automatic Problem Solving System

As plenty of countries pay more attention on AI,the more and more scientist do the research about AI and machine learning,and the more and more companies have started to focus on application of AI and machine learning.There are less related application about education which is due to the technology of AI and education that is combined not perfectly.Nowadays some newly AI algorithm technology has not been applied into the related application although the Internet is popular and people is used to searching the internet,while the research of related technology is not enough.This research is mainly to apply rule engine and computing engine technology and also contain AI and so on algorithm to automatically solve junior math questions,in the end,the system print the total answers.At the same time,the Automatically Solve Junior Math Questions(ASJMQ)system also try to solve the difficult questions which is based on parameter space and machine learning algorithm in order to find out the relationship about parameters and create knowledge that i want.What’s more,what i have researched promotes the development of AI and machine learning related technology.I have done the following research and implementation.1.The algorithm research based on parameter space in order to find out relationshipThe usual rule engine and compute engine could solve the majority of junior math questions,but the other math questions are difficult to solve.In order to so lve the kind of math questions.This system uses the maple as the symbol calculation tool.Nonpolynomial functions and semialgebraic set problems could be solved partly,such as find out function character and solve inequalities and equations.The usual symbol calculation tools can’t solve these problems,so in order to solve these problems,I use parameter space theory of Manifold learning and promote it in junior math.It is necessary to use parameter space and use the data in the space for find out the relationship in parameters and knowledge.Under the finite combination principle,i use the Cartesian of product in parameter values to get potential knowledge.These potential knowledge need to be verified and finally find out relationships of the hidden variables.2.Knowledge representation in automatic solving system of elementary mathematicsKnowledge representation is the foundation of all things over the elementary mathematics.If there is no wonderful knowledge representation in system,the goal of system about automatic solving elementary mathematics is tender,and it is impossible to take next step.Knowledge contains conclusion knowledge and common knowledge.The common knowledge contains the known knowledge and the knowledge created by the system;The conclusion knowledge concludes the target conclusion.What’s more,when the system has solved the target knowledge that can match the conclusion knowledge,which could activate the stopping rule,and the system could print the solving path that is answer.The system implemented the function that the conclusion and known knowledge could exchange.The function could enforce the ability of the system,and add the ability of creating some new mathematical problem.3.Application of HVRDPS algorithm in elementary mathematics automatic solving system.This thesis realizes the application of HVRDPS algorithm in elementart mathematics automatic solving system.A large number of methods have been written in the system and been packaged.Various types of solve operation classes have been established and HVRDPS has been embedded into various solution algorithms.And the rules that match the various solving algorithms are written and combined with various solving methods.At the same time,some design patterns are used to make the knowledge representation,method set and rule set have lower redundancy,lower coupling degree and improve system stability.