- مبلغ: ۸۶,۰۰۰ تومان
- مبلغ: ۹۱,۰۰۰ تومان
Associative classification (AC) is a new, effective supervised learning approach that aims to predict unseen instances. AC effectively integrates association rule mining and classification, and produces more accurate results than other traditional data mining classification algorithms. In this paper, we propose a new AC algorithm called the Fast Associative Classification Algorithm (FACA). We investigate our proposed algorithm against four well-known AC algorithms (CBA, CMAR, MCAR, and ECAR) on real-world phishing datasets. The bases of the investigation in our experiments are classification accuracy and the F1 evaluation measures. The results indicate that FACA is very successful with regard to the F1 evaluation measure compared with the other four well-known algorithms (CBA, CMAR, MCAR, and ECAR). The FACA also outperformed the other four AC algorithms with regard to the accuracy evaluation measure.
Phishing is an attempt, usually made through fake websites or emails, to steal an individual’s private information. Phishing websites are a significant problem, preventing users from carrying out activities via the internet. This paper aims to develop a new, fast AC algorithm called FACA and investigate FACA against four well-known AC algorithms (CBA, CMAR, MCAR, and ECAR) with regard to classification accuracy and F1 evaluation measures toward the phishing dataset. The results demonstrate that the FACA algorithm outperformed all AC algorithms with regard to accuracy and F1. Our findings also indicate thatthere is potentialfor the use of computerized data mining techniques in predicting the complex problem of phishing websites.