Research On The Application Of Performance Evaluation Based On Improved Decision Tree Algorithm

With the progress of science and technology,enterprises are paying more and more attention to talent resources.Therefore,all kinds of enterprises value performance evaluation as an effective criterion for measuring talents.so how to improve performance evaluation methods has become a problem that must be considered.Now,there are kinds of methods of performance evaluation,such as critical incident method,graphic rating scale method,paired comparison method,management by objectives method,alternately sorting method,key indicators method,behavioral anchored rating method,balanced scorecard method,360 degree performance evaluation method and many other methods.However,these methods only get some information on the surface of the data,and do not dig out the implicit correlation between the data,so they have not achieved ideal results.Therefore,the exploration of new technologies has become particularly important.Data mining technology can well dig out the deep implicit correlation between the data,make up for the shortcomings of traditional methods,and the technology has been widely applied in many fields,also worked very well.So data mining methods will be introduced to the application of performance evaluation,excavating the relationship between the performance evaluation and the influencing factors.Data classification is an important subject of data mining,now there are many kinds of data classification methods,and in these methods,because of decision tree classification algorithm has the characteristics of simple theory,quick calculation,accurate classification,easy to understand,and easy to convert the generated decision tree into classification rules,it is widely studied and applied.Therefore,we applies the decision tree classification algorithm to the performance evaluation.In many decision tree classification algorithms,ID3 algorithm and FuzzyID3 algorithm are the based algorithm of clear decision tree and fuzzy decision tree.Through the analysis and comparison of the typical decision tree algorithms,discovered that these two algorithms have obvious advantages compared with other algorithms,and can output simple rules,making it easy for users to understand.Therefore,I decides to use ID3 algorithm and FuzzyID3 algorithm as the algorithm basis of the research.And based on this,improved the computational complexity.In this paper,the concept of decision coordination degree is introduced,then combines decision coordination degree with ID3 algorithm and FuzzyID3 algorithm,which effectively reduces the computational complexity of the algorithm and realizes the improvement of the two algorithms.Here,the advantages of the improved method in the time of building a decision tree and test accuracy are proved by the classic example.Finally in the applications of X company employee performance evaluation,respectively,using the original algorithm and the improved algorithm to classify the target data,through the analysis of the classification results,further proved that the improved algorithm not only has all the advantages of the original algorithm,but also has achieved better results in the time of building a decision tree and est accuracy.In addition,according to the characteristics of the collected data of performance measurement,the operation of data fuzzy processing is added in the data preprocessing stage,through the analysis and comparison of the fuzzy data construction decision tree model build with final date and fuzzy data,found that the fuzzy decision tree model build with fuzzy data has higher test accuracy,proved the data fuzzy processing is necessary.