Empirical Analysis of film and TV play risk assessment based on text mining technology

• 1、Beijing University of Posts and Telecommunications Automation School,Beijing 100876
• 2、Who's Who Culture Industry Group, Beijing 100120

Abstract：Behind of the surface thriving film and television play market, a lot good and bad scripts filled with viewing space, only a handful of classic plays, causing this situation is the immediate screenplay preliminary evaluation mode. In this paper, based on the traditional film and television scripts subjective assessment methods, to establish the combination of subjective and objective evaluation screenplay assessment of the evaluation system, provide a new assessment method screenplay assessment. Objective assessment using the text mining technology applied to the play of the quantitative evaluation, structured analysis to the film and television scripts, get feature vector, and the dimension of the vector drop by principal component analysis (PCA). Using k nearest neighbor, support vector machines (SVM), and naive Bayesian methods to modeling the feature vectors, to select the best model for the quantitative risk assessment of screenplay. Empirical analysis the script has been broadcast on risk assessments, the accuracy of the selected model (knn)reaches 90%, to the before of broadcast television script provides practical guidance assessments.

Keywords： text mining risk assessment k nearest neighbor screenplay