Classifying Email as Spam or Non-Spam.
| Samples total |
|
| Dimensionality |
|
| Features |
|
| Targets |
|
| Task(s) |
|
The "spam" concept is diverse: advertisements for products/web sites, make money fast schemes, chain letters, pornography...
Our collection of spam e-mails came from our postmaster and individuals who had filed spam. Our collection of non-spam e-mails came from filed work and personal e-mails, and hence the word 'george' and the area code '650' are indicators of non-spam. These are useful when constructing a personalized spam filter. One would either have to blind such non-spam indicators or get a very wide collection of non-spam to generate a general purpose spam filter.
Determine whether a given email is spam or not.
~7% misclassification error. False positives (marking good mail as spam) are very undesirable.If we insist on zero false positives in the training/testing set, 20-25% of the spam passed through the filter.
Downloaded from the UCI Machine Learning Repository on March 23, 2018.
Cranor, Lorrie Faith, and Brian A. LaMacchia. "Spam!." Communications of the ACM 41.8 (1998): 74-83.
yellowbrick.datasets.loaders.load_spam