To gauge the efficiency of the model proposed in this article in precise eventualities, this article chooses Ok-means, KNN, logistic regression (LR), and random forest (RF) to check experiments with the tactic used in this text. When the gap is less than 2 m, the impact of the logistic regression algorithm will not be excellent; the typical accuracy charge is simply 68.13%, the accuracy distinction between Ok-Means, KNN, and random forest algorithm is small, and the tactic used in this article reaches 92.17%. Compared with the other four algorithms, it has the next accuracy rate. The corresponding values of the KNN method are 81.34%, 90.26%, 63.53%, and 12.235%. It may be seen that, as the space increases, the accuracy of those methods is bettering.
When the space is 2-3 m and 3-5 m, the accuracy of the OSVM algorithm is also the highest. When the space is 2-3 m, the accuracy of the logistic regression algorithm continues to be at the bottom. In the three attack eventualities of the experiment, when the space between the WiFi attacker and the ZigBee gadget is small (that’s, less than 2 m), the accuracy of K-means, KNN, logistic regression, and random forest algorithms are all beneath 90%, of which logistic regression and the accuracy of the OSVM algorithm differ by 24%. The accuracy of the other three algorithms differs from that of the OSVM algorithm by greater than 10%. When the distance is 2-3 m and 3-5 m, the OSVM algorithm also performs larger than the other four algorithms.
To gauge the performance of the proposed scheme in the actual scenarios, we simulated ten assault locations (i.e., places 11-20). Furthermore, we implemented the spoofing monitoring program, wherein the distance between the official ZigBee gadget and the WiFi attacker was lower than 2 m, 2-3 m, and 3-5 m, respectively. The OSVM algorithm used in this paper reaches 95.38%. As the distance increases to 3-5 m, the accuracy of the 5 algorithms improves. This article Evaluating the standard deviation of the accuracy of the five algorithms, it’s found that the standard deviation of the accuracy of the OSVM algorithm is at all times smaller than the other four algorithms, which signifies the detection efficiency of the OSVM is more stable than the opposite 4 algorithms.
