FredZhang7 commited on
Commit
f58c35e
1 Parent(s): 2191ac2

add categories

Browse files
Files changed (1) hide show
  1. README.md +7 -11
README.md CHANGED
@@ -45,17 +45,6 @@ language:
45
  - multilingual
46
  size_categories:
47
  - 1M<n<10M
48
- configs:
49
- - config_name: default
50
- data_files:
51
- - split: train
52
- pattern:
53
- - "phishing_features_train.csv"
54
- - "phishing_url_train.csv"
55
- - split: test
56
- pattern:
57
- - "phishing_features_val.csv"
58
- - "phishing_url_val.csv"
59
  ---
60
 
61
 
@@ -65,6 +54,12 @@ To extract features from a website, simply passed the URL and label to `collect_
65
  In the *features* dataset, there're 911,180 websites online at the time of data collection. The plots below show the regression line and correlation coefficients of 22+ features extracted and whether the URL is malicious.
66
  If we could plot the lifespan of URLs, we could see that the oldest website has been online since Nov 7th, 2008, while the most recent phishing websites appeared as late as July 10th, 2023.
67
 
 
 
 
 
 
 
68
  Here are two images showing the correlation coefficient and correlation of determination between predictor values and the target value `is_malicious`.
69
 
70
  ![Correlation Coefficient](https://i.imgur.com/LLD3pmt.png)
@@ -100,6 +95,7 @@ Let's exmain the correlations one by one and cross out any unreasonable or insig
100
  | page_rank_decimal | |
101
 
102
 
 
103
  ## Pre-training Ideas
104
 
105
  For training, I split the classification task into two stages in anticipation of the limited availability of online phishing websites due to their short lifespan, as well as the possibility that research done on phishing is not up-to-date:
 
45
  - multilingual
46
  size_categories:
47
  - 1M<n<10M
 
 
 
 
 
 
 
 
 
 
 
48
  ---
49
 
50
 
 
54
  In the *features* dataset, there're 911,180 websites online at the time of data collection. The plots below show the regression line and correlation coefficients of 22+ features extracted and whether the URL is malicious.
55
  If we could plot the lifespan of URLs, we could see that the oldest website has been online since Nov 7th, 2008, while the most recent phishing websites appeared as late as July 10th, 2023.
56
 
57
+ ## Malicious URL Categories
58
+ - Defacement
59
+ - Malware
60
+ - Phishing
61
+
62
+ ## Data Analysis
63
  Here are two images showing the correlation coefficient and correlation of determination between predictor values and the target value `is_malicious`.
64
 
65
  ![Correlation Coefficient](https://i.imgur.com/LLD3pmt.png)
 
95
  | page_rank_decimal | |
96
 
97
 
98
+
99
  ## Pre-training Ideas
100
 
101
  For training, I split the classification task into two stages in anticipation of the limited availability of online phishing websites due to their short lifespan, as well as the possibility that research done on phishing is not up-to-date: