Research on abnormal data detection of optical fiber communication network based on data mining

[Retracted] Malicious Intrusion Data Mining Algorithm of Wireless Personal Communication Network Supported by Legal Big Data

“…On the basis of the above experimental environment and parameter design, the experiment is carried out by comparing the methods of this paper, reference [5], reference [6], and reference [7]. The experimental indexes are the precision of malicious intrusion data mining in wireless personal communication network, the time consumption of mining, and the error of malicious intrusion data extraction.…”

“…The higher the value, the better the effect of mining. Therefore, results of the experimental comparison of this method and the methods from references [5], [6], and [7], to sample wireless personal communication network malicious intrusion data mining accuracy, are shown in Figure 4. The analysis of the experimental results in Figure 4 shows that as the sample intrusion data increases, four methods have changed the accuracy of malicious intrusion data mining.…”

“…The analysis of the experimental results in Figure 4 shows that as the sample intrusion data increases, four methods have changed the accuracy of malicious intrusion data mining. When the malicious data is 800, the accuracy of this method is about 98%, about 79% of the accuracy of the method in reference [5], about 82% of the accuracy of the method in reference [6], and about 75% of the accuracy of the method in reference [7]. When the malicious data is 1600, the accuracy of this method is about 99%, about 71% of the accuracy of the method in reference [5], about 80% of the accuracy of the method in reference [6], and about 65% of the accuracy of the method in reference [7].…”

“…When the malicious data is 800, the accuracy of this method is about 98%, about 79% of the accuracy of the method in reference [5], about 82% of the accuracy of the method in reference [6], and about 75% of the accuracy of the method in reference [7]. When the malicious data is 1600, the accuracy of this method is about 99%, about 71% of the accuracy of the method in reference [5], about 80% of the accuracy of the method in reference [6], and about 65% of the accuracy of the method in reference [7]. By comparing these four methods, we can see that the accuracy of this method is the highest.…”

“…This method has high detection accuracy for intrusion data, but it is easy to be interfered by high noise data in the process of classification, resulting in poor detection results. In reference [7], a data mining-based abnormal data detection method for optical fiber communication network is proposed. The operation process of this method is relatively simple, and the clustering accuracy of feature attributes in classification is high, but the amount of data after attribute fusion is less, so it cannot be widely used.…”