Collaborative Filtering Recommendation System Application Research In Intelligent Driving Coaches

In recent years our country driving rapid growth in quantity and quality on the development trend of the combination of declining. In order to solve the related problems of the traditional driving exam mode, to adapt to the self-study after the take an examination of the transformation, for students to choose their coaches in the training process, and provide intellectual support and data support flexible training,this article will focus on different stages of driving between the students and coach to study science adaptation issues.Premise of this article research is electronic commerce system of driving training school system, namely the coach of the school shall be regarded as commodity trade in O2 O. System combines hardware and software system, in the construction of the physical model, using LBS location technology and RFID equipment combining sensor technology, through the GIS service period of data collection; In the design of system software, using the data mining and network security technologies, such as through needs analysis, and interface design, set up to the coach and the students interactive software.In the design of intelligent recommendation algorithm, aiming at the existing problem of collaborative filtering algorithm to make improvements, in allusion to the problem of the initial recommendation quality is poorer, an improved algorithm of hierarchical analysis based on Delphi, through to the user’s initial data preprocessing,the recommended at the beginning of the algorithm which have high quality; For sparse problem of collaborative filtering algorithm based on graph spectral clustering algorithm was improved, and at the same time, in the calculation of similarity matching by spectral clustering algorithm to find and recommend a similar item,improve recommendation performance of the algorithm; New user extensions in traditional collaborative filtering and difficult problems, according to the actual system requirements, the collaborative filtering recommendation mechanism based on goods, reduce the new user data processing complexity. Finally through the experiment proves that the improved algorithm on intelligent recommendation system is effective.