An improved method using fuzzy system based on hybrid boahs for phishing attack detection
A fuzzy system is a rule-based system that uses human expert's knowledge which holds the truth or false values to make a particular decision. However, it is difficult to generate fuzzy parameters manually to classify data when it comes to a very complex problem. Therefore, metaheuristic algorit...
Saved in:
| Main Author: | |
|---|---|
| Format: | Thesis |
| Language: | English |
| Published: |
2022
|
| Subjects: | |
| Online Access: | http://umpir.ump.edu.my/id/eprint/37654/1/ir.An%20improved%20method%20using%20fuzzy%20system%20based%20on%20hybrid%20boahs%20for%20phishing%20attack%20detection.pdf |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | A fuzzy system is a rule-based system that uses human expert's knowledge which holds the truth or false values to make a particular decision. However, it is difficult to generate fuzzy parameters manually to classify data when it comes to a very complex problem. Therefore, metaheuristic algorithm as the optimization method is needed to solve this issue. A small experiment was being performed to choose the best metaheuristic algorithm where seven metaheuristic algorithms were analyzed in terms of their performance measurement that including accuracy, recall, precision, and f-measure. The algorithms involved were Genetic Algorithm, Differential Evolution Algorithm, Particle Swarm Optimization, Butterfly Optimization Algorithm, Teaching-Learning-Based Optimization Algorithm, Harmony Search Algorithm and Gravitational Search Algorithm. The proposed method of this study is to cater the problems occur in fuzzy systems by using optimization method. Moreover, Butterfly Optimization Algorithm and Harmony Search Algorithm were combined as optimization method led to a new method named BOAHS. The proposed algorithm has utilized the advantages of both algorithms to balance the exploration and exploitation search process. The experiment was executed by using k-fold cross validation techniques for predicting the classification algorithm performance. Thereby, two datasets; Website Phishing Dataset and Phishing Websites Dataset were used to test the performance of the proposed method. As a result, the average accuracy value for both datasets are 98.69% and 98.80% respectively. The proposed method has proven to outperform the other methods including the standard BOA and HS algorithm. |
|---|