Abstract. In previous studies, time series model has been the most common method and fuzzy time series model is an improvement on time series model. However, the accuracy in fuzzy time series model has been inevitably affected by interval length and the partitioning method for formulating effective interval can be very difficult. In addition to time series model, the grey model has also been widely studied, but it may have bad curve-fitting effects when data show great randomness. Therefore, both fuzzy time series model and grey model have problems in accuracy for forecasting tourism flow. Based on fuzzy time series model and grey model, this paper proposed a hybrid model which has been optimized by entropy and Markov model. In this hybrid model, the weight is calculated by entropy method which has been used to balance the performance of two single models. Markov model with its stable property is used for processing data sequence with large fluctuation. The result of the experiment clearly shows that this proposed hybrid model optimized the interval partition and therefore ensured the prediction accuracy.
Question answering (QA) system provides a direct, efficient and accurate way for people to obtain information. At present, open domain QA systems such as Siri and Cortana are widely used in the general field, but they cannot meet the demand of some professional fields. This paper focuses on the background and needs of QA in the tourism field, researching the relevant technologies required for the implementation of QA system, and finally completes the construction of QA system based on the knowledge graph of tourism. The main research contents of this paper are as follows: 1.An algorithm to identify the tourism entities in questions is proposed according to the characteristics of the tourism entities. 2. Referring to the ideas of Liu et al., a convolutional neural network (CNN) model is introduced into attribute linking, but in order to improve the accuracy of attribute linking, we move the similarity calculation of questions and attributes from the outside of the model to the input layer of the model, and also introduce Attention mechanism. Integrate the technology of each module to design and implement a QA system for tourism. We experiment with the system on the constructed Xi’an tourism knowledge graph, and the results prove that the system we designed can answer the natural language questions raised by users about tourism.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.