CN-122019892-A - Tourist route intelligent recommendation method and system based on scenic spot travel fusion feature similarity

CN122019892ACN 122019892 ACN122019892 ACN 122019892ACN-122019892-A

Abstract

The invention belongs to the field of artificial intelligence, and relates to an intelligent travel route recommending method and system based on scenic spot travel fusion feature similarity. According to the method, a scenic spot resource knowledge base is constructed according to scenic spot culture and travel information, and N historical scenic spots are determined according to historical travel attention information of a user. When the number of the historical scenic spots is small, the similarity between the historical scenic spots and the destination scenic spots is directly calculated, a scenic spot planning route with high similarity is selected, when the number of the historical scenic spots is large, 3-5 core scenic spots with high user attention are screened out by adopting an improved time attenuation weighting method, the similarity between the core scenic spots and the destination scenic spots is calculated, and a scenic spot planning route is selected. If the a scenery spots contain destination boundary scenery spots and the scenery spots with higher similarity exist near the boundary, the scenery spot with the lowest similarity is replaced by the scenery spot and is recommended. The method can reduce the operation amount of the server, give consideration to the demands of adjacent scenic spots, has small route variation, saves the travel cost and is more in line with the wish of target users.

Inventors

ZHOU XIANGBING
LI XI
Ran Xiaojuan
ZHANG CHENXI
LI XIAOFENG
DU SIYUAN
HUANG LIYAO
XUE DONG

Assignees

四川旅游学院
澳门城市大学

Dates

Publication Date: 20260512
Application Date: 20260202

Claims (9)

1. The intelligent travel route recommending method based on scenic spot travel fusion feature similarity is characterized by obtaining travel fusion features according to cultural information and travel information of scenic spots, integrating the travel fusion features of different scenic spots in different areas to construct a scenic spot resource knowledge base, and comprises the following steps: S1, obtaining historical tourist attractions according to historical tourist concern information of a target user, and obtaining the tourist fusion characteristics of the historical tourist attractions from a attraction resource knowledge base; s2, acquiring the text and travel fusion characteristics of all scenic spots in the target area from a scenic spot resource knowledge base, and selecting a recommendation method according to the number of the historical scenic spots, wherein if the number of the historical scenic spots does not exceed a preset threshold, the first similarity between the historical scenic spots and the scenic spots in the target area is calculated, and all the scenic spots and the corresponding first similarity in the target area are integrated to form a first set; S3, dividing the historical tourist attractions into intention historical tourist attractions and past historical tourist attractions, acquiring network browsing records of the intention historical tourist attractions, calculating the attention degree of the intention historical tourist attractions based on an improved time attenuation weighting method, and marking the attractions with high attention degree as important intention historical tourist attractions; S4, selecting a recommended scenic spots with high similarity from the first set or the second set to form a third set, determining whether adjacent scenic spots exist according to whether the scenic spots at the boundary exist in the third set, calculating the third similarity between the historical tourist attractions and the adjacent scenic spots, and comparing the similarity between the adjacent scenic spots and the a recommended scenic spots to determine whether to update the third set; s5, carrying out travel route planning on the a recommended scenic spots in the third set according to the dynamic limiting conditions of the target user.
2. The intelligent travel route recommending method based on scenic spot travel fusion feature similarity according to claim 1, wherein the construction of the scenic spot resource knowledge base comprises the following steps: S01, acquiring and integrating cultural information and tourist information of different scenic spots in different areas, wherein the cultural information comprises remote sensing image data of regional characteristics, text data and image data of national culture, and the tourist information comprises text data and image data of scenic spot evaluation; S02, adopting two dimensions The method comprises the steps of completing texture feature extraction of remote sensing image data by a filter algorithm, completing image feature extraction by adopting image2vec, and obtaining a triple representation form of scenic spot travel fusion features after multi-mode feature extraction is completed, wherein the triple representation form comprises the following steps: ; Representing scenic spots I represents the ith scenic spot, i is a positive integer; Texture features, text features, and image features representing the attraction; s03, converting the triple expression form into a cascade vector, and acquiring a compact expression of the text-to-travel fusion characteristic by using a PCA dimension reduction algorithm: ; s04, the travel fusion characteristics of all scenic spots in a region are used as a set to be expressed, and the scenic spot fusion characteristics are stored into a scenic spot resource knowledge base according to the sets of different regions.
3. The intelligent travel route recommending method based on scenic spot travel fusion feature similarity according to claim 1, wherein the number of historical travel spots is recorded as N, and the number of all scenic spots in the target area is recorded as M; In the step S2, N first similarities exist in each target area scenic spot, the largest first similarity S 1 is selected, the target area scenic spot M i and the largest first similarity S 1mi corresponding to the target area scenic spot are stored as a data set (M i ,s 1mi ), the data sets of all scenic spots in the target area are integrated, the data sets of all scenic spots in the target area are ordered according to the first similarity, a first set is formed, M i represents the ith target area scenic spot, i=1, 2, 3.
4. The intelligent travel route recommending method based on scenic spot travel fusion feature similarity as set forth in claim 3, wherein in the step S3, N second similarities exist in each scenic spot of the target area, the largest second similarity S 2 is selected, the scenic spot m i of the target area and the largest second similarity S 2mi corresponding to the scenic spot m i of the target area are stored as a data set (m i ,s 2mi ), the data sets of all scenic spots of the target area are integrated, and the data sets of all scenic spots of the target area are ordered according to the second similarities to form a second set.
5. The intelligent travel route recommending method based on the scenic spot travel fusion feature similarity according to claim 1, wherein the web browsing records comprise browsing frequency, browsing duration, access depth and interaction behavior, and the method for calculating the attention score of the intention historical tourist spots in the step S3 is as follows: s31, acquiring network browsing record data of all intention historical tourist attractions in one period, and carrying out normalization processing on the network browsing record data; S32, calculating the attention score of the behavior in the j day of the i-th intention historical tourist attraction : =ω 1 *PL+ω 2 *JS+ω 3 *HD+ω 4 *SD; PL is the sight spot browsing frequency in a day, ω 1 is the corresponding weight, JS is the average sight spot browsing time in a day, ω 2 is the corresponding weight, HD is the sight spot interaction rate in a day, ω 3 is the corresponding weight, SD is the sight spot access depth in a day, ω 4 is the corresponding weight, ω 1 +ω 2 +ω 3 +ω 4 =1;i=1,2,……,n 1 ,n 1 represents the number of intention history tourist attractions, j=1, 2, L represents the network browsing record of the target user in L days in one period; S33, determining the attention score of the ith intention historical tourist attraction : ; Lambda is the decay factor and t i is the time difference of the j-th day from the current time, day.
6. The intelligent travel route recommending method based on scenic spot travel fusion feature similarity according to claim 1, wherein the scenic spot at the boundary refers to a scenic spot with a geographic intersection between an adjacent area with a radius r and an adjacent area with the scenic spot as a center; The step S4 comprises the following specific steps: S41, selecting a recommended scenery spots with high similarity from the first set or the second set, and sorting data sets of the a recommended scenery spots according to the similarity to form a third set; S42, judging whether scenic spots at the boundary exist in the third set: If no scenic spot exists at the boundary, the third set is not updated; If the scenic spot at the boundary exists, the step S43 is entered; s43, judging whether neighboring scenery spots exist in the scenery spot neighboring area D at the boundary position: if no neighboring scenic spots exist, the third set is not updated; if the scenic spots in the adjacent areas exist, calculating a third similarity between the historical tourist attractions and the scenic spots in the adjacent areas, and judging the size of the third similarity: If the third similarity is greater than the reference similarity and the number of the reference scenic spots is only 1, replacing the data set of the reference scenic spots in the third set with an array of scenic spots in the adjacent area, and updating the third set; If the third similarity is greater than the reference similarity and the number of reference scene points is greater than or equal to 2, step S44 is entered; if the third similarity is less than or equal to the reference similarity, not updating the third set; s44, if the reference scenic spot is also the scenic spot at the boundary, the reference scenic spot is marked as a special scenic spot, and whether the reference scenic spot is the special scenic spot is judged: if the special scenic spots exist in the plurality of reference scenic spots, replacing the data set of the special scenic spots in the third set with an array of scenic spots in the adjacent area, and updating the third set; and if no special scenery spot exists in the plurality of reference scenery spots, replacing the reference scenery spot farthest from the scenery spot in the adjacent area with the scenery spot in the adjacent area, and updating the third set.
7. The intelligent travel route recommending method based on scenic spot travel fusion feature similarity according to claim 1, wherein the dynamic limit conditions of the target user in the step S5 comprise travel time, departure place, travel mode, scenic spot real-time opening time, reception capacity and traffic condition.
8. The intelligent travel route recommending method based on scenic spot travel fusion feature similarity according to claim 1, further comprising the steps of S6, expressing an unwanted intention of a target user on recommended scenic spots in a third set, recording the recommended scenic spots as exclusive scenic spots, assigning the similarity corresponding to the exclusive scenic spots in the first set and the second set to be 0, updating the first set and the second set according to the new similarity ranking, and entering step S4 again.
9. The intelligent travel route recommending system based on the scenic spot travel fusion feature similarity is characterized by being used for executing the intelligent travel route recommending method according to any one of claims 1 to 8, wherein the system comprises a user side, a server side and a cloud side; the user side comprises a man-machine interaction interface which is used for a target user to access the system and display and inquire the analysis and calculation result of the server side; The cloud comprises a scenic spot resource knowledge base, wherein scenic spot data of different scenic spots in different areas are stored in the scenic spot resource knowledge base, and the scenic spot data comprise scenic spot text and travel fusion characteristics, scenic spot cultural information and travel information; The server side comprises a first similarity calculation module, a second similarity calculation module, an analysis processing module and a third similarity calculation module; when the number of the historical tourist attractions is smaller than or equal to a preset threshold value, calculating the similarity between the historical tourist attractions and all attractions in the target area by using a first similarity calculation module; When the number of the historical tourist attractions is larger than a preset threshold value, the second similarity calculation module is utilized to focus on the similarity among the historical tourist attractions, the historical tourist attractions and all attractions in the target area; When the third set has scenic spots at the boundary, and the scenic spots in the adjacent area of the scenic spots at the boundary have scenic spots in the adjacent area, a third similarity calculation module is utilized to calculate the third similarity between the historical tourist attraction and the scenic spots in the adjacent area; The analysis processing module acquires scenic spot information from the cloud and transmits the scenic spot information to the first similarity calculation module and the second similarity calculation module, a first set, a second set and a third set are constructed according to calculation results of the first similarity calculation module and the second similarity calculation module, reference similarity and reference scenic spots are selected from the third set, similarity between scenic spots in adjacent areas and all scenic spots in the third set are compared to determine whether the third set is updated or not, and travel route planning is conducted on recommended scenic spots in the third set according to dynamic limiting conditions of target users.

Description

Tourist route intelligent recommendation method and system based on scenic spot travel fusion feature similarity Technical Field The invention belongs to the technical field of travel route recommendation, and particularly relates to an intelligent travel route recommendation method and system based on scenic spot travel fusion feature similarity. Background With the improvement of the economic level of people and the continuous progress of the big data networking technology, the travel industry develops rapidly, and the travel resource recommending software becomes an important tool for improving the travel experience of users and promoting the travel consumption. In the face of a great deal of travel information and diversified user demands, how to provide personalized and accurate travel resource recommendation for target users is particularly important. At present, the travel resource recommendation method comprises a content-based recommendation method, a user-based recommendation method, a large model recommendation method, a comprehensive recommendation method and the like, wherein the content-based recommendation method is used for researching historical behaviors of a target user, analyzing similarity between a historical travel spot and other travel spots, and recommending spots similar to the historical travel spot according to a similarity size system. If the number of the historical tourist attractions of the target user is too large, the similarity between the historical tourist attractions and other tourist attractions is calculated in sequence, the system operation amount is large, the system is required to have higher requirements and better hardware, and the running cost is likely to be increased. The existing recommendation algorithm is used for analyzing the similarity between all scenic spots in a destination area and historical tourist attractions according to a certain destination area, so that scenic spot resources in the destination area are recommended. However, in practical application, the recommended scenic spot is just located at the boundary of the destination area, and at the same time, a scenic spot is just located in the adjacent area, the scenic spot distance between the scenic spot and the boundary is relatively close, and the similarity between the scenic spot in the adjacent area and the historical tourist scenic spot of the target user is very high, so that the scenic spot belongs to the scenic spot which the target user wants to go. The existing method does not consider the problem, and the recommendation result does not fully consider the user wish and is not comprehensive. Disclosure of Invention In order to solve the technical problems, the invention provides an intelligent travel route recommending method and system based on scenic spot travel fusion feature similarity. The intelligent travel route recommending method based on scenic spot travel fusion feature similarity obtains travel fusion features according to cultural information and travel information of scenic spots, integrates the travel fusion features of different scenic spots in different areas and constructs a scenic spot resource knowledge base, and comprises the following steps: S1, obtaining historical tourist attractions according to historical tourist concern information of a target user, and obtaining the tourist fusion characteristics of the historical tourist attractions from a attraction resource knowledge base; S2, acquiring the text and travel fusion characteristics of all scenic spots in the target area from a scenic spot resource knowledge base, and selecting a recommendation method according to the number of the historical scenic spots, wherein if the number of the historical scenic spots is smaller than or equal to a preset threshold value, the first similarity between the historical scenic spots and the scenic spots in the target area is calculated, and all the scenic spots and the corresponding first similarity in the target area are integrated to form a first set; s3, dividing the historical tourist attractions into intention historical tourist attractions and past historical tourist attractions according to the condition whether the target user has visited or not; acquiring a network browsing record of the intention historical tourist attraction, calculating the attention degree of the intention historical tourist attraction based on an improved time attenuation weighting method, and marking the attraction with high attention degree as a key intention historical tourist attraction; S4, selecting a recommended scenic spots with high similarity from the first set or the second set to form a third set, determining whether adjacent scenic spots exist according to whether the scenic spots at the boundary exist in the third set, calculating the third similarity between the historical tourist attractions and the adjacent scenic spots, and comparing the similarity between the adjacent scenic