Specific Personal Alias Withdrawal from Web and Clustering of Similar Web Documents

There are many names available for a person, place or an entity on the web. If accurate alias of a particular individual is identified it becomes very useful in numerous web related tasks like information extraction, relation extraction, biomedical fields, sentiment analysis, personal name disambiguation, etc. Here, one method is projected based on referential ambiguity to find the correct alias for a given name. After accepting real name as input lexical patterns are achieved from the web. Candidate aliases are extracted with the help of these patterns. The candidate aliases are ranked using various ranking scores like co occurrence frequency, web dice, hub discounting, and degree distribution. This method improves the recall and attains a statistically considerable mean reciprocal rank. Using candidate aliases and data files, related web documents are bunched or grouped. Grouping achieves high accuracy and reduces the complexity.