Posting online reviews and rating their satisfaction on purchased products has become an increasingly popular way to share the information for anonymous candidates who has interest in purchasing the product. In addition, people leave their interests and near-future purchasing plan on the web such as search history and search query volume. From this phenomenon, the prediction of sales performance is possible in many products by mining the data sets which are left on the web by consumers' online activities. In this paper, we focused on the movie ticket sales which word-of-mouth effect is prominent, and our goal is to forecast the sales performance of the near-weekend using box-office data and external factors such as online reviews, star ratings and search volume. For this work, we gather 1.7 million online reviews and movie ratings, and we also gather the daily search volume of movies' title for past three years. Using machine learning techniques and linear modeling, we develop a model for high-accuracy predicting of ticket sales on near-future. We also analyze a relationship between ticket sales performance on weekends and box-office data, online reviews, star ratings, and search volume. Through this work, we support to decide the ideal number of screens for a given weekend, thus it contributes to a substantial increase in the rate of profit on movie markets.