The celebrity profiling is used to predict the sub-profiles like gender, fame, birth-year and occupation of a celebrity for a given textual content. The task of celebrity profiling is introduced in PAN Competition 2019. Most of the researchers in the competition have shown interest on stylistic features to differentiate the writing styles of the celebrities. In this work, a sub-profile based weighted approach is proposed to improve the accuracy of celebrity profiling. In this approach, most frequent terms are used to compute the document weight. The document weights were used to represent the document vectors instead of weights of features. The document vectors forwarded to machine learning algorithms to build the training model. The proposed method achieved competitive accuracies of 77.13% for gender prediction, 87.76% for fame prediction and 91.54% for occupation prediction. The accuracies of the proposed approach for sub-profiles prediction outperform several existing approaches for celebrity profiling.