Based On The Shortest Distance Clustering K Nearest Neighbor Classifier Research And Application

Teachers and students can not be separated from the textbooks. It plays a key role in the whole teaching work. At present, the country is vigorously promoting textbooks’composing and publications, and emerge a large number of various and plentiful textbooks. At the same time, the textbook’s market has showed uneven phenomenon, such as the form of some textbooks is simple, lacking of related resources, can not meet the economic and social developmental demanding for talents training and so on. Therefore, how to select appropriate textbooks for teaching has become the most important problem which teachers and teaching administrators should pay attention to. They will choose more suitable textbooks with the help of building up scientific and reasonable evaluating system, and teaching will be carried out more effective, teaching quality and cultivating ability will be enhance.However, how to establish a scientific and reasonable evaluation method? Only through strict and accurate data processing, calculation and experimental analysis ensure the accuracy and science of evaluation results. Through comprehensive analysis and research, data classification can recognize the quality grade of textbooks. At the same time, in order to improve the accuracy and efficiency of classification results, the traditional K nearest neighbor classification is improved, At first the material samples should be clustered by the shortest system clustering method, and then small clusters or isolated points will be classified using the K nearest neighbor method, and ultimately the quality grade of textbooks can be predicted.This paper mainly discusses the research background and significance of data mining technology, focuses on the characteristics of data classification algorithms and the existing problems, combing with the implementation of data classification techniques used in this study, relating to the application of teaching quality evaluation, a suitable algorithm will be chosen for the classification of textbook samples. Aiming at higher computational overhead problem of the K nearest neighbor (KNN) classification method, put forward clustering of samples using the shortest system clustering method first, then classify the clustering of small clusters or isolated points. This method greatly reduces the size and number of classification samples, and reduces calculating cost, then further improves the classification efficiency.The textbook data is research object of this paper, and the classification of textbook quality grade will be researched from three aspects of theory, algorithm and application. The main research results include:First of all, the common classification algorithms are introduced, then the classification of KNN is puts forward by analyzing classification algorithms’advantage and weaknesses, and this paper summarizes the existing problems of the traditional KNN classification. By studying the measures of improving the traditional KNN classification, this paper put forward clustering by the shortest system clustering method first and then classifying. So the proposed method can greatly improve the efficiency of classification.Secondly, the new methods will be used in textbook quality evaluation. Through the classification of the material sample, the quality grade of textbooks can be predicted, and then providing the scientific basis for the selection of textbooks.Finally, the superiority of the new method will be demonstrated by the experimental analysis scientifically. The performance efficiency of new method will be proved from the aspects of key parameters and data size, then the power of new method will be exhibited.