Box Office Prediction Analysis Based On Online Word-of-mouth

With the rapid development of today's society,movies not only penetrate into every aspect of people's lives,but also play an important role in national economic and culture soft power.With the industrialization and marketization of the film industry,film is no longer limited to art,but an industry and a product.Nowadays,films not only need to face creative challenges,but also need to face great marketing pressure like other commodities.Especially,with the strong support of national policies,China's film market has shown a prosperous development,which has triggered a film investment boom and made more competitive.Under such a background,the box office,as an important indicator to measure the success of a film,began to become the target of investors and producers.Online Word-of-mouth contains consumers' evaluation of products and services.It has become an important factor influencing consumers' purchase decisions by virtue of its fast propagation speed and wide propagation range,and has had an important influence on the box office.Therefore,building a box office prediction model based on the factors of Online Word-of-mouth has very important practical significance for both Online Word-of-mouth marketing and box office prediction.Take the quantity and valence of Word-of-mouth as the measurement index of Online Word-of-mouth.The affective information in Online Word-of-mouth extracted by emotion analysis method to quantify the Word-of-mouth valence index,and the unstructured data is converted into structured data that can be used for modeling.Taking Online Word-of-mouth and early box office income as explanatory variables,and the current box office income as explained variables,the econometric method was used to construct the box office prediction model.40 domestic films released in mainland China from January 2016 to April 2018 were selected for empirical analysis,20 of which were randomly selected as training sets,and the parameters of the analysis model were selected.The other 20 were used as test sets to verify model parameter selection.The results show that the box office of the previous week had a better prediction effect,and the quantity and valence of Word-of-mouth of the previous day's comments had a better prediction effect.Finally,the validity of the constructed box office prediction model is illustrated by comparing it with the autoregressive model and the autoregressive model of Word-of-mouth valence is introduced.The results show that the accuracy of prediction model in this paper is higher.At the same time,the prediction model for the box office can also be used to predict the sales of other products after a little improvement.