We propose a new topic model for analyzing purchase data with price information. Price is an important factor in consumer purchase behavior. The proposed model assumes that a topic has its own price distributions for each item as well as an item distribution. The topic proportions, which represent a user’s purchase tendency, are influenced by the user’s purchased items and their prices. By estimating the mean and the variance of the price for each topic, the proposed model can cluster related items taking their price ranges into consideration. We present its efficient inference procedure based on collapsed Gibbs sampling. Experiments on real purchase data demonstrate the effectiveness of the proposed model.We propose a new topic model for analyzing purchase data with price information. Price is an important factor in consumer purchase behavior. The proposed model assumes that a topic has its own price distributions for each item as well as an item distribution. The topic proportions, which represent a user’s purchase tendency, are influenced by the user’s purchased items and their prices. By estimating the mean and the variance of the price for each topic, the proposed model can cluster related items taking their price ranges into consideration. We present its efficient inference procedure based on collapsed Gibbs sampling. Experiments on real purchase data demonstrate the effectiveness of the proposed model.
© 2001-2024 Fundación Dialnet · Todos los derechos reservados