Clustering Based On Particle Swarm Optimization Algorithm And Its Application In The Students' Grade Management

Data mining is one of the most active branches of database studying, which analyzes data and acquires knowledge from database and data warehouse with the method of artificial intelligence. Data mining is a multidisciplinary research field. There are a lot of methods about data mining, such as clustering algorithm . Clustering analyses as an independent tool can get the data distribution, observe every clustering characteristic and focus on the further analysis to specific cluster.Particle swarm optimization (PSO) is an optimization algorithm based on iteration, which was proposed by Kennedy and Eberhart in 1995. The system that with a set of random solutions, and the best solution is to be found through a way of iteration. However ,clustering algorithm may be refrained from partial superior solution for some initial center. So the improved clustering algorithm based on particle swarm optimization is proposed in the paper. PSO is similar to the Genetic Algorithm(GA). However, PSO has no crossover and mutation, and the coding is simpler than that of GA. PSO is more effective than GA in most cases, so it is very significant.Grading the student score is an important part of education management. Traditional method of grading is based on original score, which has some weaknesses. For example, if the test questions are hard in some test, the score will be lower in the whole. Namely, the result of traditional method of grading is not reasonable and impartial. So the thought of clustering in data mining is used for reference, which can overcome the shortcoming of traditional method effectively. So this paper grades the students'score by the method of data mining and the PSO.The work is done in this paper as follows:(1) Studying and improving the clustering algorithm in data mining. Because the result of k-means clustering algorithm depends on a choice of initial clustering center, k-means clustering algorithm may be refrained from partial superior solution for some initial center. Based on the weakness resulted from the sensitivity of k-means algorithm to random center, improved tactic based on the radius is proposed in the paper.(2) Compound the particle swarm optimization and k-means, the improved clustering algorithm based on particle swarm optimization is proposed in the paper. PSO is an optimization tool based on iteration. The system starts with a set of random solutions, and the best solution is found through a way of iteration. But k-means is easy to sink into the partial superior solution . Improved clustering based on particle swarm optimization algorithm is proposed to get a much better result.(3) Regarding the weakness of traditional method of grading, and the direct k-means sinking into the partial superior solution easily, improved clustering based on particle swarm optimization algorithm is applied in grading the students'scores. It can be observed that if this method is applied in the scores of students, we can get reasonable result. Further more, it can be applied in teaching appraisal to improve teaching appraisal and serve teaching quality.