Frequent Itemset based Model for Video Quality Assessment
Frequent pattern mining is traditional method for knowledge discovery. Machine learning and knowledge discovery are inspiring many engineering solutions. Emerging technologies in multimedia have prompted a demand for quality analysis and quantification of quality evaluation algorithms. Even though many video quality metrics are proposed to automatically evaluate the quality of video, individual quality metrics fail to comprehensively quantify the quality of video. Hence advanced objective video quality analysis aims to predict quality of video based on many individual quality metrics with the opinion score provided by the observer. Machine learning is helping the field of video quality assessment with its new dimensions. In this paper, association rule-based feature selection using apriori method for video quality assessment is implemented. Evaluation of the efficacy of the prediction models is done using accuracy and found that Naive Base classifier to predict video quality excellent results than with SVM(Support Vector Machine) and KNN(K Nearest Neighbourhood).