A personalized travel route planning method considering multiple themes

Abstract

The application provides a kind of personalized travel route planning method considering multiple themes, belongs to traffic planning design technical field.The technical problems of not considering travel theme or single travel theme in travel route are solved.The technical scheme is as follows: including the following steps: S1, obtaining user travel data and scenic spot data;S2, calculating user travel preferences, and dividing scenic spot themes;S3, calculating the score of each user to each scenic spot;S4, establishing a multi-theme constraint optimization model;S5, using two-stage method and genetic algorithm to solve the model.The beneficial effects of the application are as follows: the application further meets the theme demand on the basis of considering user preferences, and sets time limit and other conditions, to provide personalized travel route suggestions for users, so as to improve the user satisfaction with travel route.

Description

Technical Field

[0001] This invention relates to the field of transportation planning and design technology, and in particular to a method for planning personalized travel routes that takes into account multiple themes. Background Technology

[0002] Themed tourism is gaining popularity, demonstrating that personalized travel itinerary planning is no longer limited to traditional sightseeing but focuses more on tourists' individual interests and needs. Therefore, how to develop multi-themed personalized travel itineraries has become a hot topic.

[0003] Research on personalized travel route planning mainly focuses on how to accurately obtain user preferences and meet single-theme or even thematic needs. One approach, "Design and Implementation of a Personalized Tourism Recommendation System Based on the Data Mining and Collaborative Filtering Algorithm," calculates user similarity and recommends attractions visited by people with high similarity. Another approach, "A Picture-Based Approach to Tourism Recommendation System," uses image-based questionnaires to implicitly guide users' preferences for tourism products, accurately recommending products to travelers with different tastes. A comprehensive survey on travel recommender systems provides separate recommendations for restaurants, museums, and hotels based on user preferences. Currently, most personalized travel route planning relies on a large number of users to improve the accuracy of identifying user preferences, but struggles to cope with situations with limited user data. Furthermore, when considering thematic needs, some approaches either ignore thematic needs or only consider specific categories under a single theme, such as recommending only restaurants. Summary of the Invention

[0004] The purpose of this invention is to provide a personalized travel route planning method that considers multiple themes. Based on review text data, it uses the LDA model and TF-IDF algorithm to calculate user preferences and thematic characteristics of attractions. In assessing user satisfaction with attractions, it not only considers the matching degree between user preferences and attraction features but also uses the entropy weight method to calculate fixed scores under indicators such as the number of returning visitors, local ratings, and the number of visitors in the past 30 days. In the optimization model, with the goal of maximizing user satisfaction, it establishes constraints on the number of themes to meet the user's given theme requirements and employs a two-stage method to simplify the model. This invention, based on a comprehensive consideration of attraction features and user preferences, further satisfies theme requirements and plans travel routes containing multiple categories.

[0005] To achieve the above-mentioned objectives, the present invention employs the following technical solution: a personalized travel route planning method considering multiple themes, comprising the following steps:

[0006] Step S1: Obtain user travel data and attraction data;

[0007] Step S2: Calculate user travel preferences and categorize attractions by theme;

[0008] Step S3: Calculate each user's rating for each attraction;

[0009] Step S4: Establish a multi-topic constraint optimization model;

[0010] Step S5: Solve the model using a two-stage method and a genetic algorithm.

[0011] Preferably, in step S1, the specific process of obtaining user travel and attraction data is as follows:

[0012] Step S11: Obtain scenic spot data through channels such as Meituan, Dianping, and Douyin, such as comprehensive ratings, evaluation datasets PD, recommended products, and the region they belong to, which can form a set of scenic spots A;

[0013] Step S12: Obtain user travel data through channels such as Meituan, Dianping, and Douyin, such as the user's region, historical rating data R, and comment text data UD;

[0014] Specifically, in step S2, the process of calculating user travel preferences and dividing attractions into themes is as follows:

[0015] Step S21: Process UD using the TF-IDF algorithm to calculate user preferences. After processing, extract s high-frequency keywords from UD to form the user preference attribute set I = <I1,I2,……,I s >, its corresponding preference value is i = <i1,i2,……,i s >, among which I s Let be the word to be calculated, and j be the comment containing the word, j∈UD. For comment j, I s Quantity, This represents the total number of words in comment j.

[0016] Step S22: Based on the scenic spot evaluation dataset PD, use the LDALDA topic model to extract sub-topics and their keywords for each scenic spot. Analyze the keyword extraction results for each scenic spot one by one, remove irrelevant words, and modify the number of each topic, but do not modify the number of keywords, so as to better reflect the actual sub-topics.

[0017] Step S23: Using the K-means clustering method, the sub-topics are clustered into n categories based on the topic words, and the relevance to each topic is calculated. The specific process is as follows:

[0018] Step S231: Randomly select the initial centers of n classes;

[0019] Step S232: In the i-th iteration, for any sample, calculate its Euclidean distance to each of the n centers based on the word frequency, and assign the sample to the class containing the center with the smallest distance.

[0020] Step S233: Update the center value of the class using methods such as the mean;

[0021] Step S234: Repeat steps S232 and S233. If the center values ​​of the n cluster centers remain unchanged, the iteration ends. Step S235: Based on the actual situation, comprehensively adjust and determine the topic labels, and further calculate the relevance of each topic word for each scenic spot in its evaluation dataset using the TD-IDF algorithm. Then, the topic label is K = <K1,K2,……,K n >, and its corresponding relevance is k = <k1,k2,……,k n >, where k n For the keyword K n Relevance.

[0022] Step S3, the specific process of calculating the score for each attraction is as follows;

[0023] Step S31, calculate the relationship between user u and attraction A i Interest matching score, formula is

[0024] Step S32: Combining local ratings, repeat visitor numbers, and the number of visitors in the past 30 days, the fixed rating for the i-th attraction is calculated using the entropy weight method. The calculation formula is r. i =∑ j w j ·a ij , where w j =(1+(lnn) -1 ·∑ i p ij ·lnp ij )·(∑ j w j ·b ij ) -1 , a ij Let b be the value of the j-th indicator for the i-th scenic spot. ij For a ijThe normalized values; specifically, step S4, the specific process of establishing the multi-topic constrained optimization model is as follows; step S41, considering the actual travel situation of users, the following assumptions are made about the model:

[0025] In step S411, if the user's travel route starts at point A1, then the set of candidate attractions is A, and the number of candidate attractions is N. A .

[0026] Step S412: If the evaluations given by each user for each attraction are genuine and reliable, and reflect their true preferences, then user u's evaluation of attraction A... i rating Where r i u For user u to visit attraction A i Interest matching score;

[0027] Step S42, with the goal of maximizing user satisfaction, the objective function can be established as follows: Where y i ={0,1} indicates whether attraction A is selected in the route. i ;

[0028] Step S43, taking into account the actual travel situation of users, the following restrictions are imposed on the model:

[0029] Step S431: The user's maximum playtime for this trip is T. max Since travel time is random, chance-constrained programming is used, and the time window constraint is: in v i For attraction A i The length of stay, t ij For attraction A i With attraction A j Travel time between;

[0030] Step S432: For the user-provided desired topic tag B, the route contains at most S max If a scenic spot does not meet the theme requirements, then the theme constraint is: in B≥1, if scenic spot A i If the topic tags meet the requirements, then K i For attraction A i collection of topic tags

[0031] Step S433: If the user's travel route must form a loop, then constraints are constructed.

[0032] Step S434: Each attraction can only be visited once, so we construct constraint (1-3):

[0033]

[0034] μ i -μ j +1≤(N A -1)(1-x ij (2)

[0035] 2≤μ i ≤N A (3)

[0036] Where k = 2, ..., N A ,μ i For A i The order of visits in the route, i = 2, ..., N A ;

[0037] In step S435, only the selected attractions can form a route, thus constraints can be constructed. Where i = 1, ..., N A ;

[0038] Specifically, step S5, the specific process of solving the model using the two-stage method is as follows:

[0039] Step S51, establishing the first-stage model and obtaining the optimal set of tourist attractions, involves the following process:

[0040] Step S511, the objective function is simplified to

[0041] Step S512, relax the constraints. Step S431 can be transformed into... The constraints in step S432 can be transformed into Where M is an infinite number and Z is an auxiliary variable.

[0042] Step S513: Solve the first-stage model using a genetic algorithm;

[0043] Step S52, the specific process of establishing the second-stage model and determining the access order is as follows:

[0044] Step S521, the objective function is simplified to

[0045] Step S522, strengthen the constraints. Step S431 can be transformed into... Step S432 can be transformed into

[0046] Step S523: Solve the second-stage model using a genetic algorithm.

[0047] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0048] 1. This invention uses comment text data and employs the LDA model and TF-IDF algorithm to accurately calculate user preferences and thematic features of each attraction. Traditional methods usually rely only on simple user ratings or recommendation systems, while this invention analyzes the thematic information hidden in the text data to make user needs more accurately matched with actual tourism resources. This method can improve the personalization of tourism planning even when individual user attributes are missing.

[0049] 2. When assessing user satisfaction with attractions, this invention incorporates multi-dimensional factors, such as the number of returning visitors, local ratings, and the number of visitors in the past 30 days. Compared to traditional single-indicator assessment methods, this comprehensive assessment model can more fully reflect the attraction's attractiveness and actual value. By calculating a fixed score using the entropy weight method, the objectivity and accuracy of the assessment are further improved, thus providing users with more reliable travel advice.

[0050] 3. In optimizing travel routes, this invention aims to maximize user satisfaction by setting constraints on the number of topics and employing a two-stage method to simplify the solution model. This optimization strategy not only considers the user's given topic requirements but also effectively avoids information overload or insufficient recommendations that may occur in travel planning by introducing constraints. By simplifying the model, the final travel route planning becomes more intuitive and practical, greatly improving user experience and satisfaction. Attached Figure Description

[0051] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used together with the embodiments of the invention to explain the invention and do not constitute a limitation thereof.

[0052] Figure 1 This is a flowchart illustrating a personalized travel route planning method that considers multiple themes, as proposed in an embodiment of the present invention.

[0053] Figure 2 This is a flowchart illustrating the calculation of user travel preferences and the segmentation of attraction keywords in the method proposed in this embodiment of the invention.

[0054] Figure 3 This is a flowchart illustrating the calculation of each user's rating for each attraction in the method proposed in this embodiment of the invention.

[0055] Figure 4 This is a flowchart illustrating the method proposed in this embodiment of the invention, which uses a two-stage approach and a genetic algorithm to solve the model. Detailed Implementation

[0056] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. Of course, the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0057] Example 1

[0058] See Figures 1 to 4 This embodiment provides a technical solution: a personalized travel route planning method considering multiple themes, comprising the following steps:

[0059] Step S1: Obtain user travel data and attraction data;

[0060] Step S2: Calculate user travel preferences and categorize attractions by theme;

[0061] Step S3: Calculate each user's rating for each attraction;

[0062] Step S4: Establish a multi-topic constraint optimization model;

[0063] Step S5: Solve the model using a two-stage method and a genetic algorithm.

[0064] Preferably, in step S1, the specific process of obtaining user travel and attraction data is as follows:

[0065] Step S11: Obtain attraction data for Xianju City through the Dianping website. Attraction types include restaurants, shopping malls, and nearby tourist attractions; in this embodiment, restaurants are the primary attractions, forming set A. The data includes comprehensive ratings, evaluation dataset PD, and recommended products.

[0066] Step S12: Obtain travel data of Dianping users through web crawling, such as their region, historical user rating data R, and comment text data UD.

[0067] Specifically, in step S2, the process of calculating user travel preferences and dividing attractions into themes is as follows:

[0068] Step S21: Use the jieba word segmentation package in Python to process the user comment text data UD, and use the TF-IDF algorithm to calculate the word frequency. After processing by the algorithm, extract s high-frequency keywords from the comment text data UD to form the user preference attribute set I = <I1,I2,……,I s >, its corresponding preference value is i = <i1,i2,……,i s >. Among them, I s Let be the word to be calculated, and j be the comment containing the word, j∈UD. For comment j, Is Quantity, This represents the total number of words in comment j.

[0069] Step S22: Taking restaurants as an example, use jieba to segment the restaurant evaluation dataset PD, and use the LDA topic model to extract sub-topics and their topic words for each restaurant. Analyze the extraction results one by one, remove irrelevant words, and modify the number of individual topics, but do not modify the number of topic words.

[0070] Step S23: Using the K-means clustering method, the sub-topics are clustered into 4 categories based on the topic words, and the relevance to each topic is calculated. The specific process is as follows:

[0071] Step S231: Randomly select the initial centers of 4 classes;

[0072] Step S232: In the i-th iteration, for any sample, calculate its Euclidean distance to the four centers based on the word frequency, and assign the sample to the class containing the center with the smallest distance.

[0073] Step S233: Update the center value of the class using methods such as the mean;

[0074] Step S234: Repeat steps S232 and S233. If the center values ​​of the four cluster centers remain unchanged, the iteration ends.

[0075] Step S235: Based on the actual situation, comprehensively adjust and determine the topic tags, which are "Environment," "Service," "Taste," "Average Price," and "Gathering," etc. Further calculate the relevance of each topic keyword for each attraction in its evaluation dataset using the TF-IDF algorithm. Then, the topic tag K = <K1,K2,……,K n >, and its corresponding relevance is k = <k1,k2,……,k n >, where k n For the keyword K n Relevance.

[0076] Step S3, the specific process of calculating the ratings for each restaurant is as follows;

[0077] Step S31, calculate the interest matching score between user u and restaurant A1, using the formula:

[0078] Step S32: Combining three indicators—local customer ratings, number of repeat customers, and number of people who have checked in within the last 30 days—the entropy weight method is used to calculate...

[0079] The fixed rating for the i-th restaurant is calculated using the formula r. i =∑ j w j ·a ij, where w j =(1+(lnn) -1 ·∑ i p ij ·lnp ij )·(∑ j w j ·b ij ) -1 , a ij Let b be the value of the j-th indicator for the i-th restaurant. ij For a ij The normalized value, k = (ln n) -1 .

[0080] Specifically, step S4, the specific process of establishing the multi-topic constraint optimization model is as follows:

[0081] Step S41, considering the actual travel situation of users, make the following assumptions about the model:

[0082] In step S411, if the user's travel route starts at point A1, then the set of candidate restaurants is A, and the number of candidate attractions is N. A .

[0083] Step S412: If the reviews given by each user for each restaurant are genuine and reliable, and reflect their true preferences, then user u's opinion on restaurant A... i rating Where r i u For user u to restaurant A i Interest matching score;

[0084] Step S413: If the user's travel route starts at point A1, then the set of candidate attractions is A, and the number of restaurants is 200. A =201.

[0085] Step S42, with the goal of maximizing user satisfaction, the objective function can be established as follows: Where y i ={0,1} indicates whether restaurant A is selected in the route. i

[0086] Step S43, taking into account the actual travel situation of users, the following restrictions are imposed on the model:

[0087] Step S431: The user's maximum playtime for this trip is T. max =12h, since travel time is random, chance-constrained programming is used, and the time window constraint is: in v i For restaurant A iThe length of stay at Restaurant A. i With Restaurant A j Using latitude and longitude as input, the distance is obtained by calling the Gaode API, and the travel time t can be calculated using average vehicle speed as the standard. ij ;

[0088] Step S432: Given the user's desired theme tag B = {"team building", "party}, and the route contains at most two restaurants that do not meet the theme requirements, then the theme constraint is: in B≥1, if restaurant A i If the topic tags meet the requirements, then K i For restaurant A i collection of topic tags

[0089] Step S433: If the user's travel route must form a loop, then constraints are constructed.

[0090] In step S434, since each attraction can only be visited once, the following constraints (1-3) are established:

[0091]

[0092] μ i -μ j +1≤200(1-x ij (2)

[0093] 2≤μ i ≤201, (3)

[0094] Where k = 2, ..., 201, μ i For A i The order of visits in the route, i = 2, ..., 201;

[0095] In step S435, only the selected attractions can form a route, thus constraints can be constructed. Where i = 1, ..., 201;

[0096] Specifically, step S5, the specific process of solving the model using the two-stage method is as follows:

[0097] Step S51, establishing the first-stage model and obtaining the optimal set of tourist attractions, involves the following process:

[0098] Step S511, the objective function is simplified to

[0099] Step S512, relax the constraints. Step S431 can be transformed into... The constraints in step S432 can be transformed into Where M is an infinite number and Z is an auxiliary variable.

[0100] Step S513: Solve the first-stage model using a genetic algorithm;

[0101] Step S52, the specific process of establishing the second-stage model and determining the access order is as follows:

[0102] Step S521, the objective function is simplified to

[0103] Step S522, strengthen the constraints. Step S431 can be transformed into... Step S432 can be transformed into

[0104] Step S523: Solve the second-stage model using a genetic algorithm.

[0105] Example 2

[0106] Based on the personalized travel route planning method that considers multiple themes in the aforementioned embodiments, the following is a specific example:

[0107] Step S1: Using web scraping technology, obtain the review data of 20 users and the data of 200 restaurants and 50 scenic spots on the Dianping website. Some of the review data of the restaurant "Wei Xian Ju" is shown in Table 1.

[0108] Table 1. Partial evaluation data of the restaurant "Wei Xian Ju"

[0109]

[0110] Step S2 involves calculating user travel preferences and categorizing restaurants and scenic spots using thematic tags. After removing irrelevant high-frequency words, the keywords are "service," "price," "taste," "environment," "parking," "dining," and "value for money."

[0111] Step S3: Combine the interest matching scores of 20 users with those of 200 restaurants, and the fixed scores of the restaurants to calculate the overall user rating for each restaurant. Using the entropy weight method, the weights of the three indicators—local resident rating, number of repeat customers, and number of people who have checked in within the last 30 days—are 0.4, 0.3, and 0.3, respectively. The scores obtained by this method and the scores after plain text topic matching are compared with the actual user ratings; some results are shown in Table 2.

[0112] Table 2 Partial Scoring Results

[0113] 1 5 4.6 4 2 5 4.8 4.5 3 4.5 4.3 3.5 4 4 4.1 3.5 5 5 4.7 4 6 3.5 4.0 3 7 4 4.1 3.5 8 3.5 4 2

[0114] Calculations show that the overall accuracy of the ratings obtained by this method reaches 80%, with accuracy rates of 88.2% for the worst and 87.5% for the best restaurants. Compared to traditional methods, this method effectively identifies user preferences by utilizing text reviews, even when individual user privacy information (age, gender, preferences, etc.) is lacking; while the accuracy rate of pure text matching methods is only 73%. This indicates that the fixed restaurant ratings considered in this method, based on the matching degree of user preferences and product attributes, are beneficial to further improving the accuracy.

[0115] Step S4: Given the user's travel theme "team building" or "gathering", and the starting point, a multi-theme constraint optimization model is established based on constraints such as travel time.

[0116] Step S5: Using a two-stage method and a genetic algorithm, the optimal route is calculated for five different starting points. The results are shown in Table 3. Taking Baita Station as an example, the optimal route under the multi-theme requirement constraint is "Baita Station - Potan Ancient Town - Shanyin Coffee (Danzhu Branch) - Shenxianju Scenic Area - Xiaoshan Village Farmhouse - Baita Station," which satisfies the theme requirements of "team building" and "gathering," with scores of 4.8, 4.4, 4.1, and 4.7 respectively, an average score of 4.5, and a total travel time of 65 minutes. However, if the theme requirement is not considered, the average score of the optimal route is slightly higher than 4.5, but only one route satisfies the theme requirement. Therefore, it can be concluded that the optimal solution under the multi-theme constraint, at the cost of slightly lower scores, can basically meet the theme requirements, which aligns with the objectives of this invention.

[0117] Table 3 Comparison of Schemes

[0118]

[0119]

[0120] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A personalized travel route planning method considering multiple themes, characterized in that, Includes the following steps: S1, obtain user travel data and attraction data; S2 calculates user travel preferences and categorizes attractions by theme; In step S2, user travel preferences are calculated and attraction themes are categorized, specifically including the following steps: Step S21: Use the jieba word segmentation package in Python to process the user comment text data. The TF-IDF algorithm was used to calculate word frequencies, and the results were processed from the comment text data. Extract High-frequency keywords constitute a set of user preference attributes. Its corresponding preference value ,in, , For the words that need to be calculated, For comments containing words, , For comments In Quantity, For comments Total number of words in the text; Step S22, evaluate the restaurant's data set. Word segmentation was performed, and the LDA topic model was used to extract the sub-topics and their keywords for each restaurant. The extraction results were analyzed one by one, irrelevant words were removed, and the number of each topic was modified, but the number of keywords was not modified. Step S23: The sub-topics are clustered into 4 categories based on the topic words using the K-means clustering method, and the relevance with each topic is calculated. S3 calculates each user's rating for each attraction; S4, Establish a multi-topic constraint optimization model; Step S4, establishing a multi-topic constraint optimization model, specifically includes the following steps; Step S41, considering the actual travel situation of users, make the following assumptions about the model: Step S411, the starting point of the user's travel route The candidate attraction set is The number of candidate attractions is ; Step S412: If the user's evaluation of each attraction is authentic and reliable, and reflects their true preferences, then the user... For tourist attractions rating ,in For users For tourist attractions Interest matching score; Step S42, with the goal of maximizing user satisfaction, establish the objective function as follows: ,in Indicates whether attractions are selected in the route. ; Step S43: Considering the actual travel situation of users, impose restrictions on the model; In step S43, considering the actual travel situation of users, the following restrictions are imposed on the model: Step S431, the user's maximum playtime for this trip is Since travel time is random, chance-constrained programming is used, and the time window constraint is: ,in , For tourist attractions The length of stay For tourist attractions With attractions Travel time between; Step S432, based on the user's desired topic tags The route contains at most If a scenic spot does not meet the theme requirements, then the theme constraint is: ,in If the scenic spot If the topic tags meet the requirements, then , For tourist attractions collection of topic tags ; Step S433: If the user's travel route must form a loop, then constraints are constructed. ; In step S434, since each attraction can only be visited once, the following constraint is established: ,（1）； ，（2）； ，（3）； in , for The order of visits in the route, ; Step S435: If only the selected attractions constitute the route, then constraints are constructed. i in which ; S5 uses a two-stage method and a genetic algorithm to solve the model; Step S5, which uses a two-stage method to solve the model, specifically includes the following steps; Step S51, establishing the first-stage model and obtaining the optimal set of tourist attractions, involves the following process: Step S511, the objective function is simplified to ; Step S512, relax the constraints; step S431 can be transformed into The constraints in step S432 can be transformed into ,in, It is an infinitely large number. As an auxiliary variable; Step S513: Solve the first-stage model using a genetic algorithm; Step S52, establishing the second-stage model and determining the access order: Step S521, the objective function is simplified to ; Step S522, strengthen the constraints; step S431 can be transformed into... Step S432 can be transformed into ; Step S523: Solve the second-stage model using a genetic algorithm.

2. The personalized travel route planning method considering multiple themes according to claim 1, characterized in that, Step S1, in which user travel and attraction data are obtained, specifically includes the following steps: Step S11: Obtain attraction data through Meituan, Dianping, and Douyin channels, and combine the rating and evaluation datasets. Recommended products and their respective regions constitute a collection of attractions. ; Step S12: Obtain user travel data, including the region and historical user rating data, through Meituan, Dianping, and Douyin. and comment text data .

3. The personalized travel route planning method considering multiple themes according to claim 1, characterized in that, Step S23 involves using the K-means clustering method to cluster subtopics into four categories based on keyword information, and calculating the relevance to each subtopic. The specific process is as follows: Step S231: Randomly select the initial centers of 4 classes; Step S232, in the In each iteration, for any sample, the Euclidean distance from its word frequency to the four centers is calculated, and the sample is assigned to the class of the center with the smallest distance. Step S233: Update the center value of the class using the mean method; Step S234: Repeat steps S232 and S233. If the center values ​​of the four cluster centers remain unchanged, the iteration ends. Step S235: Based on the actual situation, comprehensively adjust and determine the topic tags, which are "Environment," "Service," "Taste," "Average Price," and "Gathering." Calculate the relevance of each topic keyword for each attraction in its evaluation dataset using the TF-IDF algorithm. Then, the topic tags are... The corresponding relevance is ,in Keywords Relevance.

4. The personalized travel route planning method considering multiple themes according to claim 1, characterized in that, In step S3, the score for each attraction is calculated, which specifically includes the following steps; Step S31, calculate user With attractions Interest matching score, formula is ; Step S32: Combining local ratings, repeat visitor numbers, and the number of visitors in the past 30 days, the entropy weight method is used to calculate the... The fixed rating for each attraction is calculated using the following formula: ,in , , For the first The first attraction The values ​​of each indicator, for The normalized value.

Images