Many of search engines are habitual to access these personal names aliases for betterment of search. Perfect recognition of existing article is valuable in diverse web related tasks like sentiment analysis, information reclamation, personal name disambiguation, and relation extraction. With the growth of web data many users try to share their knowledge over Internet. Various methods are proposed to personal name bipolarization from the web and web related data. Variety of scheme is also proposed for data mining and information extraction apart from web text analysis. Although the person name bipolarization can be predicted using PCA, but here bipolarization of person name can be predicted using kernel based PCA so that the proposed methodology provides high accuracy as compared to the existing PCA technique.