Research On Tourism Suitability Evaluation Model Based On Time Series Data

At present,tourism is the main economic activity for entertainment.With the continuous increase in the number of tourists,how to accurately improve the quality of tourism should be the focus of attention.The quality of tourism during the trip is affected by many factors.Changes in climatic factors,development of scenic spots,satisfaction,and the length of comfort period will affect tourists' travel and choice of tourist destinations.Based on other scholars' research on tourism climate comfort,this paper takes Changbai Mountain as the research background,takes meteorological dimension data as the focus of tourism suitability prediction,and integrates other multi-dimensional related elements to achieve a more comprehensive and realistic approach.prediction results.The main research contents of this paper include:(1)Combining the current development status and research background of time series data,the research of time series data forecasting is divided into traditional methods and machine learning forecasting methods;Combined with the research background of tourism,it focuses on how to predict the suitability of tourism through algorithms,and how to provide theoretical support for rational planning and improvement of tourism-related departments and enterprises through the prediction results.(2)In the prediction of periodic Time series data,traditional models ignore the problem of long-term dependence of data.This paper introduces the idea of Dynamic programming in DTW(Dynamic Time Warping)algorithm and calculates the internal relationship between the interval before and after periodic data.Repeated use of historical data with high similarity in scale changes of time series data is realized.Combined with LSTM(Long Short-Term Memory)model for weighted fusion prediction,an improved DTW-LSTM model was proposed,in which Adam optimization method was used to update the weight of LSTM network,and early stop method was used to further reduce the probability of model overfitting.By setting up different comparative experiments to prove the validity of the model;(3)Aiming at the problem of single dimension in the existing tourism suitability prediction research,through the analysis of influencing factors,this paper takes the weather-related characteristics as the main dimension,and combines the relevant characteristics of tourism information and scenic spot information as auxiliary dimensions to jointly construct a multi-dimensional tourism Suitability assessment model(Multidimensional tourism suitability model,MDTW-LSTM).The model is analyzed and optimized from the aspects of feature selection,data processing,model parameters,etc.By further designing the weighted fusion principle,rationally assigning weights,and merging the prediction results of the three dimensions,the prediction accuracy is further improved.Optimization plays a positive role in excavating urban tourism value and enhancing public tourism experience.At the same time,it also plays an auxiliary role in improving tourist satisfaction and promoting the formulation of better tourism urban planning,which has certain application potential.