<![CDATA[A film is one of the entertainment media in human life. It makes the film industry becomes a potential business to get promising income. Prediction of success of a film financially supports film producers, production houses, distributors, and stakeholders in making decisions. The existing research on film predictions is from a popularity point of view. There is also research to predict success from a financial point of view, but using social media film rating data, which has been released. This paper is a study to find a predictive model for the success of films from the financial aspect of films that have not been released. The prediction model uses the Backpropgation neural network (BNN) algorithm by calculating the accuracy, precision, recall, and f-score. The observed variables are a combination of film success parameters, namely, budget, genre, production company, holiday, runtime, and competition factor. The study results show that the use of more recent data, although less data, can produce better accuracy values, compared to using more data, films released in older years. From a financial point of view, the difference in the year the film was released, especially if it was tens of years, significantly affects the prediction results. Finally, the number of hidden layers and hidden nodes has a significant non-linear role in determining the prediction results.]]>
Copyrights © 2020
