Game Price Recommendations Using Neural Networks

Thesis Type Bachelor
Thesis Status
Student Nikita Grimm
Thesis Supervisor

Playing video games is limited by monetary constraints for a large amount of gamers. The typical selling price of a game is at $60 and many users tend to buy games they are interested in only when the games have reached a lower price. In this work, we survey how much users are willing to pay for specific games. In other words, how they perceive the value of the given games. We then use this data, combine it with attributes from the games such as its genres, themes and platforms, to train a neural network. We then use this neural network to give purchase price recommendations for other games. The price recommendations are supposed to tell the users the prices at which they will feel satisfied purchasing the games. This should be useful in cases where users haven’t collected enough information on a game yet to decide a satisfactory price themselves. We determine that roughly half of the recommendations are confirmed by the users.