Thesis Análisis y predicción del éxito de un videojuego mediante minería de opinión en reseñas
Loading...
Date
2024
Authors
Journal Title
Journal ISSN
Volume Title
Program
Ingeniería Civil Informática
Departament
Campus
Campus Santiago San Joaquín
Abstract
La abundancia de la oferta de videojuegos en el mercado dificulta la búsqueda del juego para los jugadores y también obstaculiza el camino al éxito para los desarrolladores independientes, que no cuentan con medios de promoción de sus juegos y cuyos juegos pueden pasar desapercibidos. La existencia de plataformas de juegos en línea como STEAM y el alto número de reseñas en ellas pueden proporcionar información valiosa para predecir el éxito de un videojuego o incluso para predecir los aspectos más interesantes para los jugadores.
El presente trabajo emplea la minería de opinión para evaluar el éxito de un videojuego asignando a cada oración de una reseña una valoración positiva, neutra o negativa. A través de varias iteraciones, se logró un algoritmo no supervisado de extracción y clasificación de aspectos.
Para validar los resultados de esta clasificación, se contrastaron los números de aspectos positivos y negativos con el éxito de 25 juegos de 3 géneros distintos, logrando comprobar la existencias de correlación positiva con aspectos positivos y negativa con aspectos negativos. El factor de mayor correlación con el éxito del juego fue el ratio entre reseñas positivas y negativas en el mes anterior al ranking obtenido.
The abundance of video game offerings on the market makes it difficult for gamers to find a game and also hinders the path to success for independent developers, who have no means of promoting their games and whose games may go unnoticed. The existence of online gaming platforms such as STEAM and the high number of reviews on them can provide valuable information to predict the success of a video game or even to predict the most interesting aspects for gamers. The present work employs opinion mining to evaluate the success of a video game by iv assigning each sentence of a review a positive, neutral or negative rating. Through several iterations, an unsupervised aspect extraction and classification algorithm was achieved. To validate the results of this classification, the numbers of positive and negative aspects were contrasted with the success of 25 games of 3 different genres, managing to verify the existence of positive correlation with positive aspects and negative correlation with negative aspects. The factor with the highest correlation with the success of the game was the ratio between positive and negative reviews in the month prior to the obtained ranking.
The abundance of video game offerings on the market makes it difficult for gamers to find a game and also hinders the path to success for independent developers, who have no means of promoting their games and whose games may go unnoticed. The existence of online gaming platforms such as STEAM and the high number of reviews on them can provide valuable information to predict the success of a video game or even to predict the most interesting aspects for gamers. The present work employs opinion mining to evaluate the success of a video game by iv assigning each sentence of a review a positive, neutral or negative rating. Through several iterations, an unsupervised aspect extraction and classification algorithm was achieved. To validate the results of this classification, the numbers of positive and negative aspects were contrasted with the success of 25 games of 3 different genres, managing to verify the existence of positive correlation with positive aspects and negative correlation with negative aspects. The factor with the highest correlation with the success of the game was the ratio between positive and negative reviews in the month prior to the obtained ranking.
Description
Keywords
Mineria de opinión, Videojuegos, STEAM, Análisis de sentimientos
