Skripsi
KLASIFIKASI MALICIOUS URL PADA FILE MENGGUNAKAN METODE K-NEAREST NEIGHBOR BERDASARKAN LEXICAL FEATURE EXTRACTION.
Various forms of attack models ranging from hosting, spreading malware and phishing websites, these actions can start from accessing the Uniform Resource Locator (URL) or files that contain malicious links in them. A Uniform Resource Locator (URL) is a special identifier used to find resources over the internet. Malicious URLs can be a threat using various types of attacks, usually malicious URLs are disguised so they are easy to miss. What needs to be done to differentiate between malicious URLs and normal URLs is to use feature extraction to identify various important characteristics of malicious URLs. The extraction feature used is a lexical feature which consists of 18 features. After extraction, the results of the unbalanced dataset will be resampled using oversampling with SMOTE. This research uses the k-nearest neighbor machine learning algorithm to classify the dataset. K-Nearest Neighbor is a different characteristic classification algorithm that determines the class to which unlabeled data belongs using distance to calculate the nearest neighbors. This algorithm is able to achieve high classification accuracy and provide the best results. This research obtained evaluation results with the highest accuracy value of 98.78%, precision 98.785%, recall 98.795% and f1-score 98.79%.
Inventory Code | Barcode | Call Number | Location | Status |
---|---|---|---|---|
2307006257 | T130126 | T1301262023 | Central Library (Referens) | Available |
No other version available