Development of an Effective Machine Learning Model to Optimize Ad Placements in AVOD using Divergent Feature Extraction Process and Adaboost Technique

Background of the study: The Advertisement-based Video-on-Demand (AVoD) platform gets benefits of each advertiser who can place their ads. Due to its dynamic context, the current static optimization methodologies become inadequate. Methods: This research work has developed the machine learning approach for effective Ad placement in the AVOD. Initially, audio and video frames are collected from the standard data resource. From the audio frame, Mel-Frequency Cepstral Coefficients (MFCC) and spectral feature is extracted. Similarly, the motion feature and Histogram of Oriented Gradients (HOG) feature are extracted from the video frame. The extracted features are combined and subjected to Principal Component Analysis (PCA) for reducing the dimension. The PCA feature is further sent to the Adaboost network for effective video placement. Results: The finding of the developed Adaboost model attains 93.0 in terms of accuracy. Conclusion: This analysis shows the developed model provide an optimal solution when compared with baseline techniques.