Publications

Abstract : Each year, the quantity of released films increases. Several of the films have flopped at the box office, while others have been box office hits. The researchers developed a model for predicting the success or failure of forthcoming films based on a range of variables. The factors include the date and month of release, the cast and director, the narrative and genre, and the language. Prior to the film’s release, the suggested model will forecast its success or failure. The researchers employed robustly optimized BERT pretraining approach (RoBERTa) embeddings and a neural network model to estimate movie ratings. RoBERTa is referring to the model of the sentence transformer. It is capable of bidirectional text data reading, guaranteeing that the sentence’s meaning is correct. The researchers employed RoBERTa embedding and input the output of RoBERTa into a neural network model that predicted a value between 0 and 10 (rating). This concept will benefit film manufacturing firms since manufacturers incur significant losses as a result of investment. They will gain from this technique. This will assist the business in earning as much profit as feasible. It has been demonstrated through experiments that the neural network model outperforms the performances obtained by combining the RoBERTa embedding method with neural network and other machine learning models.