Research Of Classification Of The Relative Motion Problem In Elementary Mathematics Based On SVM

Since nearly seven years of development experience in artificial intelligence fifties of the last century, there are already quite good results. Applied Maths answer machine may achieve precisely built on fruitful predecessors. Machine answer Applied Maths need to go through the subject read, understand topics, answer machine and human-output step of this series. Title name suggests understanding of the subject application text classification and information extraction method of input that you first need a large number of topics in the text among certain types of topic and other categories of classification, and then extract the direct relationship between the title and statement implied relations, classification good topic by the specific problem-solving frame corresponding to solve, title extracted the number of various relationships are filled to problem-solving framework. This article is the work done by the subject classification, specifically, the relative movement of this specific type of topic focused identified from the title text. To achieve this goal, the paper constructed a mathematical topic text classifier, the classifier based on SVM algorithm, through classified training sample topic construct forecasting models to mathematical subject classification prediction sample set to achieve a classification relative motion problems.Support vector machine (Support Vector Machine, SVM) theory is a complete, adaptable, global optimization, generalization good performance classifier. SVM little over-fitting, for linearly inseparable data set or dimension of feature vectors high datasets correct classification rate is relatively high, suitable for text data classification. Because of this, on the basis of learning SVM attached background theory above, from the title text categorization reality, choose to build a math problem SVM text classifier.The main research work is as follows:First, this paper study a specific problem of implicit relationship extraction while solving mathematics problems, that is whether to add an relative motion equation when machine solving mathematics problems, which means to recognize if the application problem is a relative motion problem or not. Using words of bag method to represent extracted features after doing Chinese word segmentation to the input train set of mathematics problem text. Then input test set when training classification model has been done, using the same method to do the feature vector representation. Finally SVM classification algorithm to recognize whether the problems in test set are relative motion problems.Second, the paper collected mathematics test set from the junior high school authoritative teaching materials, to do empirical research through experimental verification of the performance of the classification model. Experiment results show that the classification model can be 100% rate to recognize elementary mathematics application problem whether is relative motion application problems.