Microsoft adCenter Labs has interesting demos of algorithms they are using to help target their ads. Google obviously has similar tools, but they have kept them internal (probably to keep spammers and SEO guessing). I ran through some of them for this site. Results follow. Content Categorization:
Categories Confidence |
—|—
| Computing\Software | 0.110
—|—
People & Chat\Homepages | 0.063
Computing\Internet | 0.061
Entertainment\Games | 0.050
Computing\Computer Science | 0.050
Computing\Networks & Comm. | 0.036
Computing\Sales | 0.032
People & Chat\Email | 0.029
People & Chat\Forums & Lists | 0.027
Computing\Multimedia | 0.025
Entertainment\Humor & Fun | 0.024
Computing\Hardware | 0.023
People & Chat\Chat | 0.022
Entertainment\Music | 0.020
Computing | 0.01
Demographics Prediction: Gender: Male Oriented Age: 25~34 Oriented Online Commercial Intention: Result: NonCommercial (Page) Probabilities for Each OCI Type: NonCommercial Prob.: 0.91448 Commercial- Informational Prob.: 7.6531e-002 Commercial-Transactional Prob.: 8.9882e-003