{[ promptMessage ]}

Bookmark it

{[ promptMessage ]}

hw2 - CSE 5800 Mining/Learning and the Internet-HW2 Due...

Info iconThis preview shows page 1. Sign up to view the full content.

View Full Document Right Arrow Icon
CSE 5800 Mining/Learning and the Internet—HW2 Due 6:30pm, Wed, Oct 5 Submit Server: course=cse5800 , project=hw2 Implement and evaluate LERAD (LEarning Rules for Anomaly Detection). Do not generate “wildcard rules” in Step 1 (in the paper) since they get relatively high scores in small data sets. 1. Allow parameters: (a) number of pairs of examples for generating candidate rules ( L in the paper) (b) maximum number of rules per pair of examples ( M in the paper) (c) number of examples in the sample set ( | S | in the paper) (d) number of examples in the validation set as a percentage of the entire training set [e.g. 10% means 90% for training, 10% for validation] 2. Vary the score threshold, report AUC (area under curve) upto 1%, 10%, and 100% false alarm rate. 3. Three data sets: (a) toy data set on the course web site (b) intrusion detection on the course web site (c) your own data set 4. A report (in pdf) that discusses the following: (a) Sensitivity analysis of parameters: for the second data set, i. vary each of the four parameters (keeping the other three constant),
Background image of page 1
This is the end of the preview. Sign up to access the rest of the document.

{[ snackBarMessage ]}