Unsupervised detection of web pages corresponding to a similarity class
A method of detecting web pages belonging to at least one similarity class from a plurality of web pages includes determining clusters of the plurality of web pages based on characteristics of the content of the web pages. For each of the determined clusters, at least one metric is determined indicative of similarity among resource locators associated with the web pages of that cluster. A determination of web pages belonging to the at least one similarity class is based on the determined clusters and the determined similarity metrics.
1. A method of detecting web pages belonging to at least one similarity class from a plurality of web pages, the method comprising:
determining clusters of the plurality of web pages based on characteristics of the content of the web pages;
determining, for each of the determined clusters, at least one metric indicative of similarity among resource locators associated with the web pages of that cluster, wherein the metric indicative of similarity among resource locators is an entropy metric; and
determining the web pages belonging to the at least one similarity class based on the determined clusters and the determined similarity metrics, including determining a cluster having web pages belonging to the at least one similarity class based on local entropy of selected keys of the resource locators for the web pages of that cluster and a statistical measure of the similarity metrics.
2. The method of claim 1 , wherein:
the step of determining clusters includes shingling.
3. The method of claim 1 , wherein:
the at least one similarity class is a plurality of similarity classes; and
the method further comprises, for each similarity class, characterizing that similarity class as distinct from the other similarity classes.
4. The method of claim 1 , further comprising:
employing a crawler to gather the plurality of web pages.
5. The method of claim 1 , wherein:
the similarity metric for a cluster is an indication of statistical distribution of the resource locators for the web pages of that cluster.
6. The method of claim 1 , wherein:
the statistical measure comprises a KL-divergence of the similarity metrics.
7. The method of claim 1 , wherein:
the similarity class is soft 404.
8. The method of claim 1 , wherein:
the similarity class is a class of web pages whose content indicates that a domain was not able to fulfill a web page request.
9. The method of claim 1 , further comprising:
indicating the web pages belonging to the at least one similarity class as belonging to the at least one similarity class.
10. The method of claim 1 , further comprising:
parsing web pages determined to belong to the at least one similarity class to identify the at least one similarity class.
11. The method of claim 1 , further comprising:
indicating the web pages belonging to the at least one similarity class as belonging to the at least one similarity class; and
analyzing the plurality of web pages at least in part in view of the indications.
12. A computer program product for detecting web pages belonging to at least one similarity class from a plurality of web pages, the computer program product comprising at least one computer-readable storage medium having computer program instructions stored therein which are operable to cause at least one computing device to:
determine clusters of the plurality of web pages based on characteristics of the content of the web pages;
determine, for each of the determined clusters, at least one metric indicative of similarity among resource locators associated with the web pages of that cluster, wherein the metric indicative of similarity among resource locators is an entropy metric; and
determine the web pages belonging to the at least one similarity class based on the determined clusters and the determined similarity metrics, including determining a cluster having web pages belonging to the at least one similarity class based on local entropy of selected keys of the resource locators for the web pages of that cluster and a statistical measure of the similarity metrics.
13. The computer program product of claim 12 , wherein:
the computer program instructions operable to cause the at least one computing device to determine clusters includes computer program instructions operable to cause the at least one computing device to perform shingling on the web page content.
14. The computer program product of claim 12 , wherein:
the at least one similarity class is a plurality of similarity classes; and
the computer program instructions are operable to cause the at least one computing device to further, for each similarity class, characterize that similarity class as distinct from the other similarity classes.
15. The computer program product of claim 12 , wherein:
the similarity metric for a cluster is an indication of statistical distribution of the resource locators for the web pages of that cluster.
16. The computer program product of claim 12 , wherein:
the similarity class is soft 404.
17. The computer program product of claim 12 , wherein:
the similarity class is a class of web pages whose content indicates that a domain was not able to fulfill a web page request.
18. The computer program product of claim 12 , wherein: the computer program instructions are operable to cause the at least one computing device to further:
parse web pages determined to belong to at least one similarity class to identify the similarity class.
19. The computer program product of claim 12 , wherein the computer program instructions are operable to cause the at least one computing device to further:
indicate the web pages belonging to the at least one similarity class as belonging to the at least one similarity class; and
analyze the plurality of web pages at least in part in view of the indications.
20. A computing system including at least one computing device, configured to detect web pages belonging to at least one similarity class from a plurality of web pages, the at least one computing device configured to:
determine clusters of the plurality of web pages based on characteristics of the content of the web pages;
determine, for each of the determined clusters, at least one metric indicative of similarity among resource locators associated with the web pages of that cluster, wherein the metric indicative of similarity among resource locators is an entropy metric; and
determine the web pages belonging to the at least one similarity class based on the determined clusters and the determined similarity metrics, including determining a cluster having web pages belonging to the at least one similarity class based on local entropy of selected keys of the resource locators for the web pages of that cluster and a statistical measure of the similarity metrics.
21. The computer program product of claim 20 , wherein:
the similarity class is soft 404.
22. The computer program product of claim 20 , wherein:
the similarity class is a class of web pages whose content indicates that a domain was not able to fulfill a web page request.