Baseline results

[amritbhanu]

Experiment

• Comparison against 7 learners: linear svm, rbf svm, LR, 8-NN, NB, RF, DR.
• 5 evaluation measures
• 25 repeats. 20% training and 80% testing.
• raw data distribution: non-phished:phished: 5:6

Results

• All Features

• Feature selections from weka: Prefix_Suffix, having_Sub_Domain, SSLfinal_State, Request_URL, URL_of_Anchor, Links_in_tags, SFH, web_traffic and Google_Index

[amritbhanu]

• Important Attributes:
  • Prefix_Suffix, having_Sub_Domain, SSLfinal_State, Request_URL, URL_of_Anchor, Links_in_tags, SFH, web_traffic and Google_Index

[amritbhanu]

Data Dictionary - https://archive.ics.uci.edu/ml/machine-learning-databases/00327/

[timm]

BTW: