Name
Title | ||
|---|---|---|
Identifying Movie Genre Compositions Using Neural Networks And Introducing Genrec - A Recommender System Based On Audience Genre Perception |
Abstract | ||
|---|---|---|
Film industries all around the world continue to produce thousands of movies every year. Each such movie is generally classified into 2-5 genres on popular movie websites and streaming platforms. However, such genre classifications are based on the nature of the movie script or movie scenes alone and do not usually provide a measurement of the percentage of genres in the movie. In this work, we present an approach to break down a movie into its respective genre compositions based on viewer opinion using neural networks and introduce GenRec - a content-based movie recommender system using these opinion-based genre metrics. To train machine learning models, we have created the world's largest movie reviews to genres dataset with the help of the popular Internet Movie Database (IMDb). We analyze our dataset with ensemble classification algorithms such as Nearest Neighbors and Naive Bayes and move on to ensemble logistic regression models in order to breakdown a movie into its genre compositions. We have also trained neural network models - DNN, CNN and RNN and found that DNN outperforms the rest with an R2-score of 0.825 and RMSE of 0.060. Finally, we designed our recommender system from thousands of movies by applying similarity algorithms on the audience perceived genre compositions of movies as obtained from the neural network. In addition, we also discuss use cases of how our genre composition system can be used for analyzing movie reception and talk about some of the future work that may be conducted with our dataset. |
Year | DOI | Venue |
|---|---|---|
2020 | 10.1109/ICCCS49678.2020.9276893 | PROCEEDINGS OF THE 2020 5TH INTERNATIONAL CONFERENCE ON COMPUTING, COMMUNICATION AND SECURITY (ICCCS-2020) |
Keywords | DocType | Citations |
neural network, recommendation system, movie genre prediction, audience perception, machine learning | Conference | 0 |
PageRank | References | Authors |
0.34 | 0 | 3 |
Name | Order | Citations | PageRank |
|---|---|---|---|
| Aditya Pal | 1 | 0 | 0.34 |
| Abhilash Barigidad | 2 | 0 | 0.34 |
| Abhijit Mustafi | 3 | 0 | 1.01 |