Research On Spam Filtring Based On Improved Dendritic Cell Algorithm

A common and daunting security issue for email users is that they receive a lot of spam every day.Currently,the traditional countermeasure in most email systems is a simple filtering mechanism that blocks or quarantines unwanted emails based on user-defined keywords.Aiming at the problems of low accuracy and recall rate and unstable filtering of traditional spam filtering technology,the dendritic cell algorithm(DCA)has been improved and applied to spam filtering.The improved algorithm is superior in terms of recall,accuracy,and false positive rate.Dendritic cell algorithms are mostly used to detect network intrusions and network anomalies,and have achieved good results in this regard.Due to the similarities between intrusion detection,spam filtering and the immune system,the researchers designed the classical dendritic cell algorithm and applied it to spam filtering,and obtained preliminary research results.However,the classical dendritic cell algorithm has certain defects.For example,the dendritic cell algorithm has too many parameters and the weights in the signal processing formula are empirical values or random values,and the signal definitions are also insufficient.These problems lead to unstable mail filtering accuracy and recall rate.In order to improve the precision and recall rate of dendritic cell algorithm filtering spam,the following improvements have been made to the dendritic cell algorithm.(1)Because most of the signals and parameters of traditional dendritic cell algorithms need to be artificially defined,and experiments show that too many parameters and signal definitions do not improve the performance of the algorithm.Therefore,the four types of input signals of the traditional dendritic cell algorithm are reduced to two types of input signals,and the three types of output signals generated by the newly defined parameters are reduced to two types of output signals.(2)For the problem that the weights in the signal processing formula are empirical values or randomly generated,the dynamic optimization algorithm is introduced to optimize the weight parameters and experimentally tested.The experimental results show that the improved algorithm improves the accuracy and recall rate of spam filtering and reduces the false positive rate,and the algorithm is stable compared with other algorithms.(3)Based on the improved dendritic cell algorithm,we propose a multi-strategy filtering model that combines bayesian algorithm,logistic regression algorithm and improved dendritic cell algorithm to form a multi-strategy spam filtering model to filter spam.The model optimizes the experimental results and improves the stability,but the model reduces the speed of the algorithm.(4)The proposed multi-strategy filtering model is used to design and test the enterprise's spam filtering system.The test results show that the designed mail filtering system can effectively filter spam.The UCI Spam Base data set was used as the experimental data set,and the improved Dendritic Cell Algorithm(IDCA)and the proposed multi-strategy filtering model were tested.The experiment shows that the improved algorithm is superior to the classical dendritic cell algorithm in detection rate and false positive rate.The improved dendritic cell algorithm has significantly improved the detection rate of spam,reduced false positive rate and improved stability.The recall rate and precision rate are 0.95 and 0.90 respectively,and the rate of false positives is also relatively stable.