Learning to estimate prices of video games
Traditional video game reviews and scoring are unable to account for personal tastes and the financial situation of a given user. This could lead to a user being dissatisfied with a game purchase if it was either too expensive for what it offered or unsuited to his tastes. To solve this problem and to give better recommendations, a neural network can be used to evaluate the price at which a game becomes worth buying to a given user. The goal of this thesis is to devise and train a neural network on a game’s properties, such as its developer, genre, release platforms and price, to achieve this goal. As target data, previously given user votes on what a fair price for a given game is will be used.