Movie Rating Prediction using Ensemble Learning Algorithms

Abstract

Over the last few decades, social media platforms have gained a lot of popularity. People of all ages, gender, and areas of life have their presence on at least one of the social platforms. The data that is generated on these platforms has been and is being used for better recommendations, marketing activities, forecasting, and predictions. Considering predictions, the movie industry worldwide produces a large number of movies per year. The success of these movies depends on various factors like budget, director, actor, etc. However, it has become a trend to predict the rating of the movie based on the data collected from social media related to the movie. This will help a number of businesses relying on the movie industry in making promotional and marketing decisions. In this report, the aim is to collect movie data from IMDB and its social media data from YouTube and Wikipedia and compare the performance of two machine learning algorithms-Random Forest and XGBoostbest known for their high accuracy with small datasets, but large feature set. The collection of data is done from multiple sources or APIs.

Sentiment Analysis of Movies Based on Natural Language Processing

“…The author focused on the user's conformity in rating with a matrix-factorization-based conformity modeling technique. The result was a significant improvement compared to RMSE, so conformity modeling is important for movie rating studies [1].…”

“…By transforming the 1-10 rating result to only three categories--"Great", "OK", and "Poor", it gave a higher accuracy. (1) In a research to predict movie ratings before the movie is released, the author used data from IMDB including IMDB score, director, gross, budget and so on to train the mode. It concluded that Random forest gave the best prediction accuracy and number of voted users, number of critics for review, numbers of Facebook likes, duration of the movie and gross collection of movie affect the score most strongly [9].…”

Prediksi Rating Film IMDb Menggunakan Decision Tree

“…Penelitian tersebut menggunakan metode kNN dan menghasilkan prediksi kesusksesan sebuah film dengan rating hits, netral dan gagal. Pada penelitian yang dilakukan oleh (Mhowwala, Sulthana dan Shetty, 2020) data dari IMDb, Youtube dan wikipedia diolah menggunakan algoritma random forest dan XGBoost untuk melakukan prediksi rating film. Dari penelitian tersebut didapat akurasi 95% pada XGBoost.…”