Scenic spot passenger flow prediction method and system based on hotel order and user preference

Abstract

The application discloses a scenic spot passenger flow prediction method and system based on hotel orders and user preferences, and the method comprises the following steps: obtaining hotel order data from different platforms; wherein the hotel order data comprises a user identifier, a hotel geographical position, a check-in time and a check-out time; determining a stay time window of the user at the destination based on the check-in time and the check-out time; wherein the number of users is multiple; performing user preference extraction and candidate scenic spot screening based on the user identifier and the hotel geographical position, and determining a final candidate scenic spot set; calculating an access tendency score of each candidate scenic spot in the final candidate scenic spot set; performing scenic spot passenger flow prediction on a target date based on the stay time window, the final candidate scenic spot set and the access tendency score, and obtaining the final predicted passenger flow of each scenic spot on the target date. The above scheme can improve the accuracy and timeliness of scenic spot passenger flow prediction.

Description

Technical Field

[0001] This application relates to the field of scenic area visitor flow prediction technology, specifically to a method and system for predicting scenic area visitor flow based on hotel orders and user preferences. Background Technology

[0002] Tourist flow forecasting is a core foundation for smart tourism management and public resource allocation. Accurate advance prediction of tourist flow provides crucial decision-making support for measures such as visitor flow control, transportation capacity allocation, security personnel deployment, and emergency resource reserves, directly impacting the quality of the tourist experience and the level of public safety in tourism.

[0003] Existing methods for predicting visitor flow at scenic spots mainly fall into two categories. The first category relies on ticket reservation data, gate card swipe records, or real-time sales data from the ticketing system. It estimates future visitor flow by statistically analyzing the number of ticket-buying users or by extrapolating time series data. The fundamental limitation of this method is that its prediction occurs after users have made their ticket purchase decisions, making it a "lagging statistic" or "short-term extrapolation." The lead time for prediction is extremely limited, failing to provide sufficient time for management to respond. The second category of methods is based on historical visitor flow data from the same period, using regression analysis, time series models, or graph neural networks for statistical prediction. This type of method relies solely on macro-historical patterns, ignoring the individual travel information and spatiotemporal distribution of potential tourists currently en route. It struggles to capture sudden fluctuations in visitor flow, resulting in coarse-grained predictions that cannot be refined to specific dates. Summary of the Invention

[0004] This application aims to provide a method and system for predicting tourist flow in scenic areas based on hotel orders and user preferences, which can improve the accuracy and timeliness of tourist flow prediction.

[0005] The technical solution of this application is implemented as follows: In a first aspect, embodiments of this application provide a method for predicting tourist flow in scenic areas based on hotel orders and user preferences, the method comprising: Acquire hotel order data from different platforms; wherein, the hotel order data includes user identifier, hotel location, check-in time and check-out time; Based on the check-in time and check-out time, a user's stay time window at the destination is determined; wherein, the number of users is multiple; Based on the user identifier and the hotel's geographical location, user preferences are extracted and candidate scenic spots are filtered to determine the final set of candidate scenic spots; Calculate the visitation preference score for each candidate scenic spot in the final candidate scenic spot set; Based on the dwell time window, the final candidate scenic spot set, and the visitor tendency score, the scenic spot visitor flow is predicted for the target date, and the final predicted visitor flow for each scenic spot on the target date is obtained.

[0006] In the above scheme, the step of extracting user preferences and filtering candidate scenic spots based on the user identifier and the hotel's geographical location to determine the final set of candidate scenic spots includes: Based on the user identifier, the user's historical scenic spot visit records are obtained; wherein, the historical scenic spot visit records include scenic spot name and scenic spot type tag; Based on the scenic area name and the scenic area type label, user preferences are extracted to obtain a user preference vector; Based on the hotel's geographical location, the user preference vector, and the spatial radius threshold, candidate scenic spots are screened to determine the final set of candidate scenic spots.

[0007] In the above scheme, the step of filtering candidate scenic spots based on the hotel's geographical location, the user preference vector, and the spatial radius threshold to determine the final set of candidate scenic spots includes: Based on the hotel's geographical location and the geographical coordinates of each scenic spot pre-stored in the scenic spot database, the spatial distance between the user's hotel and each scenic spot is calculated; and based on the spatial distance between the user's hotel and each scenic spot, as well as the spatial radius threshold, candidate scenic spots in the spatial neighborhood are filtered to obtain a set of candidate scenic spots around the user. Determine the scenic area type label vector for each scenic area in the surrounding candidate scenic area set; Based on the user preference vector and the scenic area type label vector of each scenic area, calculate the semantic matching degree of the user's preference with each scenic area; By using the semantic matching degree and preference threshold between the user and each scenic spot, the scenic spots in the surrounding candidate scenic spot set are filtered to determine the final candidate scenic spot set.

[0008] In the above scheme, calculating the visit preference score for each candidate scenic spot in the final candidate scenic spot set includes: Based on the semantic matching degree of the user's preference with each scenic spot, the semantic matching degree of the preference of each candidate scenic spot in the final candidate scenic spot set is determined; Calculate the popularity score and normalized spatial distance score of each candidate scenic spot in the final candidate scenic spot set; Based on the popularity score, the normalized score of spatial distance, and the semantic matching degree of preference for each candidate scenic spot, the visitation tendency score of each candidate scenic spot in the final candidate scenic spot set is calculated.

[0009] In the above scheme, calculating the scenic area popularity score and the normalized score of spatial distance for each candidate scenic area in the final candidate scenic area set includes: Based on the spatial distance between the user's hotel and each scenic spot, the spatial distance between each candidate scenic spot in the final candidate scenic spot set and the user's hotel is determined. Based on the spatial distance between each candidate scenic spot and the user's hotel, and the spatial radius threshold, the normalized score of the spatial distance of each candidate scenic spot in the final candidate scenic spot set is calculated. Obtain the historical average daily visitor volume of all scenic spots in the scenic spot database; and determine the global minimum and global maximum values ​​based on the historical average daily visitor volume of all scenic spots. Based on the historical average daily visitor flow of all scenic spots, the global minimum and the global maximum, normalization processing is performed to determine the popularity score of each candidate scenic spot in the final candidate scenic spot set.

[0010] In the above scheme, the step of predicting the visitor flow of scenic spots on a target date based on the stay time window, the final candidate scenic spot set, and the visitor tendency score, to obtain the final predicted visitor flow of each scenic spot on the target date, includes: Users are filtered based on the dwell time window and the target date to determine valid users; Determine the final candidate scenic spot set and access preference score of the effective user; and calculate the preference weight of each candidate scenic spot in the final candidate scenic spot set of the effective user based on the final candidate scenic spot set and access preference score. Based on the effective users and the preference weights of each candidate scenic spot, the initial predicted visitor flow of each candidate scenic spot is calculated. The initial predicted visitor flow for each candidate scenic spot is corrected to obtain the final predicted visitor flow for each scenic spot on the target date.

[0011] In the above scheme, the step of correcting the initial predicted visitor flow for each candidate scenic spot to obtain the final predicted visitor flow for each scenic spot on the target date includes: Obtain the actual total number of people with all valid orders and the historical average visitor flow for each candidate scenic spot during the same period; Based on the historical average visitor flow of each candidate scenic spot and the initial predicted visitor flow of each candidate scenic spot, the probability weight of the correction factor for each candidate scenic spot is calculated. Based on the actual total number of people and the probability weights of the correction factors for each candidate scenic spot, the random predicted flow of people for each candidate scenic spot is calculated. Based on the random predicted visitor flow of each candidate scenic spot, the initial predicted visitor flow of each candidate scenic spot is corrected to obtain the final predicted visitor flow of each scenic spot on the target date.

[0012] Secondly, embodiments of this application provide a scenic area visitor flow prediction system based on hotel orders and user preferences. This system includes: an acquisition module, a determination module, a calculation module, and a prediction module; wherein... The acquisition module is used to acquire hotel order data from different platforms; wherein, the hotel order data includes user identifier, hotel geographical location, check-in time and check-out time; The determining module is used to determine the user's stay time window at the destination based on the check-in time and the check-out time; wherein, there are multiple users; based on the user identifier and the hotel's geographical location, user preference extraction and candidate scenic spot screening are performed to determine the final candidate scenic spot set; The calculation module is used to calculate the visitation preference score for each candidate scenic spot in the final candidate scenic spot set. The prediction module is used to predict the visitor flow of scenic spots on a target date based on the stay time window, the final candidate scenic spot set, and the visitor tendency score, so as to obtain the final predicted visitor flow of each scenic spot on the target date.

[0013] Thirdly, embodiments of this application provide a scenic area visitor flow prediction device based on hotel orders and user preferences, comprising: a processor and a memory; wherein, The memory is used to store computer programs; The processor is configured to call and run the computer program from the memory to perform the method as described in the first aspect.

[0014] Fourthly, embodiments of this application provide a computer-readable storage medium storing executable instructions for causing a processor to perform the method described in the first aspect.

[0015] This application provides a method and system for predicting tourist flow at scenic spots based on hotel orders and user preferences. The method includes: acquiring hotel order data from different platforms; wherein the hotel order data includes user identifiers, hotel geographical locations, check-in time, and check-out time; determining the user's dwell time window at the destination based on the check-in time and check-out time; wherein there are multiple users; extracting user preferences and filtering candidate scenic spots based on the user identifiers and hotel geographical locations to determine a final set of candidate scenic spots; calculating the visitor preference score for each candidate scenic spot in the final set of candidate scenic spots; and predicting the tourist flow at scenic spots on a target date based on the dwell time window, the final set of candidate scenic spots, and the visitor preference score, thus obtaining the final predicted tourist flow for each scenic spot on the target date. In the above scheme, the user's dwell time window is constructed using the check-in and check-out times from the hotel order data. Based on the hotel geographical location and the dwell time window, the potential activity range of users is locked within a dual spatiotemporal boundary, realizing a spatial mapping from "hotel check-in data" to "potential scenic spot visitors." This breaks through the bottleneck of traditional methods' reliance on scenic spot ticket booking data and significantly extends the lead time for tourist flow prediction. Based on the user identifier and the hotel's geographical location, user preferences are extracted and candidate scenic spots are screened to determine the final set of candidate scenic spots. This achieves structured extraction of user preferences and accurate semantic matching of scenic spot attributes, improving the personalization and rationality of candidate scenic spot selection. The visitor preference score for each candidate scenic spot in the final set is calculated, providing a comprehensive evaluation basis with a unified metric for subsequent visitor flow probability allocation. Based on the stay time window, the final set of candidate scenic spots, and the visitor preference score, visitor flow prediction for the target date is performed, obtaining the final predicted visitor flow for each scenic spot on the target date. This allows for the output of predicted visitor flow for each scenic spot before users have booked tickets, improving the accuracy and timeliness of visitor flow prediction. Attached Figure Description

[0016] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the specification, serve to explain the technical solutions of this application. Obviously, the drawings described below are merely some embodiments of this application, and those skilled in the art can obtain other drawings based on these drawings without any inventive effort.

[0017] The flowcharts shown in the accompanying drawings are merely illustrative and do not necessarily include all content and operations / steps, nor do they necessarily have to be performed in the described order. For example, some operations / steps can be broken down, while others can be combined or partially combined; therefore, the actual execution order may change depending on the specific circumstances.

[0018] Figure 1 An optional flowchart illustrating a scenic area visitor flow prediction method based on hotel orders and user preferences, provided as an embodiment of this application; Figure 2 A schematic diagram of a scenic area visitor flow prediction system based on hotel orders and user preferences is provided for an embodiment of this application; Figure 3 This is a schematic diagram of a scenic area visitor flow prediction device based on hotel orders and user preferences, provided as an embodiment of this application. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the specific technical solutions of this application will be further described in detail below with reference to the accompanying drawings of the embodiments of this application. The following embodiments are used to illustrate this application, but are not intended to limit the scope of this application.

[0020] Unless otherwise defined, all technical and scientific terms used in this application have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains. The terminology used in this application is for the purpose of describing embodiments of this application only and is not intended to be limiting of this application.

[0021] In the following description, references to "some embodiments," "this embodiment," "this application embodiment," and examples, etc., describe a subset of all possible embodiments. However, it is understood that "some embodiments" may be the same subset or different subset of all possible embodiments and may be combined with each other without conflict.

[0022] If the application documents contain similar descriptions such as "first / second", the following explanation shall be added: In the following description, the terms "first / second / third" are used only to distinguish similar objects and do not represent a specific order of objects. It is understood that "first / second / third" may be interchanged in a specific order or sequence where permitted, so that the embodiments of this application described herein can be implemented in an order other than that illustrated or described herein.

[0023] This application provides a method for predicting tourist flow in scenic areas based on hotel orders and user preferences. Figure 1 This is an optional flowchart illustrating a scenic area visitor flow prediction method based on hotel orders and user preferences, provided as an embodiment of this application. Figure 1 The steps shown are explained.

[0024] S101. Obtain hotel order data from different platforms; among which, hotel order data includes user identifier, hotel location, check-in time and check-out time.

[0025] In some embodiments of this application, hotel order data includes user identifier, hotel location, check-in time, and check-out time.

[0026] In some embodiments of this application, a scenic area visitor flow prediction method based on hotel orders and user preferences is adapted to the scenic area visitor flow prediction scenario.

[0027] In some embodiments of this application, a method for predicting tourist flow in scenic areas based on hotel orders and user preferences is adapted to a system for predicting tourist flow in scenic areas based on hotel orders and user preferences.

[0028] In some embodiments of this application, hotel order data from users on different platforms or mini-programs is obtained through a unified interface and communication protocol. The hotel order data includes user ID (user_id[i]), hotel location (hotel_loc[i], including longitude (lon[i]) and latitude (lat[i]), check-in time (check_in[i]), and check-out time (check_out[i]). The user ID (user_id[i]) is used to associate historical data of the same user on different platforms.

[0029] S102. Based on check-in and check-out times, determine the user's stay time window at the destination; where there are multiple users.

[0030] In some embodiments of this application, user i among multiple users is used as an example for illustration. Based on the user's check-in time check_in[i] (in date) and check-out time check_out[i] (in date), the stay time window of user i at the destination is determined as stay_window[i] = [check_in[i], check_out[i]].

[0031] S103. Based on user identifiers and hotel geographical locations, extract user preferences and screen candidate scenic spots to determine the final set of candidate scenic spots.

[0032] In some embodiments of this application, S103 can be implemented by S1031-S1033, as follows: S1031. Based on the user identifier, obtain the user's historical scenic spot visit records; wherein, the historical scenic spot visit records include the scenic spot name and scenic spot type tag.

[0033] For example, the user's historical scenic spot visit records history_view[i] can be obtained by associating the user identifier user_id[i], including fields such as scenic spot name and scenic spot type label.

[0034] S1032. Based on the scenic area name and scenic area type label, extract user preferences to obtain a user preference vector.

[0035] S1033. Based on the hotel's geographical location, user preference vector, and spatial radius threshold, candidate scenic spots are screened to determine the final set of candidate scenic spots.

[0036] In some embodiments of this application, the spatial distance between the user's hotel and each scenic spot is calculated based on the hotel's geographical location and the geographical coordinates of each scenic spot pre-stored in the scenic spot database. Based on the spatial distance between the user's hotel and each scenic spot, and a spatial radius threshold, candidate scenic spots in the spatial neighborhood are filtered to obtain a set of candidate scenic spots surrounding the user. The scenic spot type label vector for each scenic spot in the set of candidate scenic spots is determined. Based on the user's preference vector and the scenic spot type label vector for each scenic spot, the semantic matching degree of the user's preference with each scenic spot is calculated. Through the semantic matching degree of the user's preference with each scenic spot and the preference threshold, the scenic spots in the set of candidate scenic spots in the surrounding area are filtered to determine the final set of candidate scenic spots.

[0037] For example, the historical scenic spot visit records of user i, history_view[i], are parsed to obtain the set of specific scenic spots visited by the user in the past. The scenic spot type label system is predefined into 8 categories: {natural scenery, historical culture, theme park, leisure and vacation, city sightseeing, religious temples, characteristic blocks, zoos and botanical gardens}, and the category index m∈[1,8]. The scenic spot database pre-stores the type label vector of each scenic spot, which is an 8-dimensional bool vector (0 / 1), allowing a scenic spot to belong to multiple categories at the same time.

[0038] To calculate the cumulative hit count of each category tag for user i in historical visits, we need to iterate through each scenic spot visited by the user and add each component of the scenic spot's category tag vector to the corresponding category count (if the same scenic spot is visited repeatedly, the count is added repeatedly). We construct a frequency vector freq_vec[i][8] = [cnt[i][1], cnt[i][2], ..., cnt[i][8]], and normalize the frequency vector to obtain the user preference vector user_pref_vec[i][8]. If there are no historical visit records, each component is uniformly assigned a value of 1 / 8.

[0039] Centered on the hotel location hotel_loc[i] where user i is staying, and with a spatial radius threshold R (default 5km), all surrounding scenic spots within a distance of R are retrieved, and a candidate scenic spot set cand_set_geo[i] is constructed around user i. Specifically: The scenic area database pre-stores the geographical coordinates of each scenic area j: scenic_loc[j] = (scenic_lon[j], scenic_lat[j]); scenic_lon[j] is the longitude of scenic area j, and scenic_lat[j] is the latitude of scenic area j. The spatial distance dist[i][j] between user i's hotel and scenic area j is calculated using spherical distance (e.g., the Haversine formula). If dist[i][j] ≤ R, then scenic area j is added to cand_set_geo[i].

[0040] Based on the obtained set of surrounding candidate scenic spots cand_set_geo[i], a second-stage preference semantic filtering is performed as follows: The scenic area type label vector of scenic area j is denoted as scenic_type_vec[j][8] (m-dimensional 0 / 1 vector, allowing multiple labels, such as a scenic area belonging to both "natural scenery" and "leisure and vacation"). The semantic matching degree of user i and scenic area j, match_score[i][j], can be calculated by the following formula: ; The preference semantic matching score, match_score[i][j], is the dot product of the user preference vector and the scenic spot type label vector. Each component in the user preference vector user_pref_vec[i][m] represents the strength of the user's preference for that category label, and scenic_type_vec[j][m] is the scenic spot type label vector for scenic spot j; m is 8-dimensional data; the value in each dimension is 0 or 1, where 1 indicates that the scenic spot possesses that category attribute, and 0 indicates that the scenic spot does not possess that category attribute. The result is essentially the sum of the product of each attribute of the scenic spot and the attributes possessed by the scenic spot, directly reflecting the degree of matching between user preferences and scenic spot attributes. A higher value indicates that the scenic spot better matches the user's preferences.

[0041] The dot product method naturally supports multi-label scenic spots—the more preference attributes a scenic spot has, the greater the potential matching opportunity, without the need for additional normalization processing.

[0042] Set a preference threshold θ=0.15. If match_score[i][j]≥θ, then scenic spot j is included in the final candidate scenic spot set cand_set[i] through semantic filtering; otherwise, it is removed.

[0043] Understandably, constraining a user's spatial activity range to the accessible area around the hotel creates a spatial mapping from "a user staying in a city" to "which nearby scenic spots the user might visit," providing a calculable set of candidate scenic spots for subsequent preference matching. By statistically analyzing and normalizing the frequency of type tags for historical scenic spots visited by users, discrete travel records are transformed into a preference vector with a unified dimension. Simultaneously, dot product operations are used to calculate the semantic matching degree between the preference vector and the scenic spot type tag vector, enabling a quantifiable and comparable matching evaluation between user preferences and scenic spot attributes.

[0044] S104. Calculate the visit preference score for each candidate scenic spot in the final candidate scenic spot set.

[0045] In some embodiments of this application, S104 can be implemented by S1041-S1043, as follows: S1041. Based on the semantic matching degree of user preferences with each scenic spot, determine the semantic matching degree of preferences for each candidate scenic spot in the final candidate scenic spot set.

[0046] S1042. Calculate the popularity score and normalized spatial distance score of each candidate scenic spot in the final candidate scenic spot set.

[0047] In some embodiments of this application, the spatial distance between each candidate scenic spot in the final candidate scenic spot set and the user's hotel is determined based on the spatial distance between the user's hotel and each scenic spot; based on the spatial distance between each candidate scenic spot and the user's hotel, and a spatial radius threshold, a normalized score for the spatial distance of each candidate scenic spot in the final candidate scenic spot set is calculated; the historical average daily visitor flow of all scenic spots in the scenic spot database is obtained; and based on the historical average daily visitor flow of all scenic spots, a global minimum and a global maximum value are determined; based on the historical average daily visitor flow of all scenic spots, the global minimum and the global maximum value, normalization processing is performed to determine the scenic spot popularity score of each candidate scenic spot in the final candidate scenic spot set.

[0048] S1043. Based on the popularity score, normalized spatial distance score, and preference semantic matching degree of each candidate scenic spot, calculate the visitation tendency score of each candidate scenic spot in the final candidate scenic spot set.

[0049] For example, for each candidate scenic spot j in the final candidate scenic spot set cand_set[i], considering the three factors of preference semantic matching degree, spatial distance and scenic spot popularity, the access preference score pref_score[i][j] of user i for scenic spot j is calculated: pref_score[i][j]=α×match_score[i][j]+β×dist_norm[i][j]+γ×hot_score[j]; Wherein, match_score[i][j] is the calculated semantic matching degree of preference (dot product value), reflecting the degree of fit between user preferences and scenic area attributes; dist_norm[i][j] is the normalized score of spatial distance, dist_norm[i][j]=1-dist[i][j] / R, with a value range of [0,1], and the closer the distance, the higher the score; hot_score[j] is the scenic area popularity score, which is based on the scenic area's historical average daily visitor flow avg_flow[j] and is normalized by Min-Max: hot_score[j]=(avg_flow[j]-min_flow) / (max_flow-min_flow), with a value range of [0,1]; α, β, and γ are weight coefficients, satisfying α+β+γ=1, with the default values ​​being α=0.5, β=0.2, and γ=0.3. The pref_score[i][j] value ranges from [0,1], and a higher value indicates a stronger overall tendency for users to visit the scenic spot; min_flow and max_flow are the global minimum and global maximum values ​​of the historical daily average visitor flow avg_flow[j] for all scenic spots in the scenic spot database, respectively. min_flow=min_j(avg_flow[j]); max_flow=max_j(avg_flow[j]); Using global Min-Max normalization makes the popularity scores of different scenic spots comparable, avoiding score bias caused by differences in the size of the scenic spots themselves. If the avg_flow[j] of all scenic spots is equal (extreme case), then all hot_score[j] are set to 0.5 to ensure numerical stability.

[0050] Understandably, by integrating the preference semantic matching degree, normalized spatial distance score, and global normalized scenic spot popularity score with configurable weight coefficients to form a unified dimensional fusion, a comprehensive tendency score with clear physical meaning is output, providing a unified measurement benchmark for passenger flow probability allocation.

[0051] S105. Based on the stay time window, the final candidate scenic spot set, and the visitor tendency score, predict the scenic spot traffic on the target date to obtain the final predicted traffic for each scenic spot on the target date.

[0052] In some embodiments of this application, the visitor preference score is normalized within the candidate scenic spots into preference weights that satisfy the normalization constraints, which are directly used as the probability of users selecting each scenic spot. This is combined with the actual number of orders to complete the probability allocation of the initial predicted visitor flow. On this basis, a correction factor probability weight is introduced to uniformly model all external, unenumerable random factors (real-time congestion, weather changes, and instantaneous interest shifts, etc.) as visitor flow offsets between scenic spots, and these are weighted and fused with the initial predicted visitor flow to form the final predicted visitor flow that takes into account both deterministic preference patterns and the influence of random disturbances.

[0053] Understandably, the visitor preference score is transformed into a probability distribution that satisfies normalization constraints, thus achieving a reasonable allocation of the initial predicted visitor flow. At the same time, a correction factor probability weight is introduced to uniformly model the unenumerable external random factors as the shift in visitor flow between scenic spots. By weighting and fusing the initial predicted visitor flow with a preset balance coefficient, the final predicted visitor flow retains the deterministic patterns based on user preferences and spatial constraints, while effectively absorbing the impact of unpredictable random disturbances, thereby more closely approximating the mixed probability distribution of the actual visitor decision.

[0054] In some embodiments of this application, S105 can be implemented by S1051-S1054, as follows: S1051. Based on the dwell time window and target date, filter users to determine valid users.

[0055] It should be noted that a valid user is some or all of the multiple users.

[0056] S1052. Determine the final candidate scenic area set and access preference score of effective users; and calculate the preference weight of each candidate scenic area in the final candidate scenic area set of effective users based on the final candidate scenic area set and access preference score.

[0057] S1053. Based on the effective users and the preference weights of each candidate scenic spot, calculate the initial predicted visitor flow for each candidate scenic spot.

[0058] S1054. Adjust the initial predicted visitor flow for each candidate scenic spot to obtain the final predicted visitor flow for each scenic spot on the target date.

[0059] In some embodiments of this application, the actual total number of people in all valid orders and the historical average visitor flow of each candidate scenic spot are obtained; based on the historical average visitor flow of each candidate scenic spot and the initial predicted visitor flow of each candidate scenic spot, the correction factor probability weight of each candidate scenic spot is calculated; based on the actual total number of people and the correction factor probability weight of each candidate scenic spot, the random predicted visitor flow of each candidate scenic spot is calculated; based on the random predicted visitor flow of each candidate scenic spot, the initial predicted visitor flow of each candidate scenic spot is corrected to obtain the final predicted visitor flow of each scenic spot on the target date.

[0060] For example, let's predict tourist traffic at a scenic area for a specific future target date (denoted as target_date). First, we filter out valid bookings that are in a checked-in state on the target date (target_date); then, we complete the prediction in two steps: Valid user filtering based on target date: Set the target date target_date. For all hotel orders i, based on the constructed stay time window stay_window[i] = [check_in[i], check_out[i]], if check_in[i] ≤ target_date ≤ check_out[i], then user i is at the destination on target_date and is marked as a valid user for subsequent statistics; otherwise, they are removed.

[0061] Initial visitor flow allocation: For each valid user i, let its final candidate scenic spot set be cand_set[i], and its visit preference score be pref_score[i][j]. The visit preference score is normalized into preference weights within the candidate set: ; Wherein, weight[i][j] is the preference weight; K is the number of candidate scenic spots for effective user i; pref_score[i][j] is the access preference score of effective user i for scenic spot j; cand_set[i] is the final candidate scenic spot set; and pref_score[i][k] is the access preference score of effective user i for scenic spot k.

[0062] Since this application assumes that each user can only visit one scenic spot per day, weight[i][j] represents the probability that user i will choose scenic spot j on target_date. The initial predicted visitor flow for scenic spot j is calculated as follows: ; Where group_size[i] is the actual number of orders for valid user i (which can be obtained from the order data, and the default value is 1); max_user is the total number of valid users.

[0063] Inter-scenic area visitor flow correction based on random selection probability: Actual visitor decisions are influenced by many random factors (such as real-time crowd density, scenic area accessibility, weather changes, social media popularity, and instantaneous shifts in user interests), which cannot be fully enumerated in offline preference models. This application uniformly models these random factors as inter-scenic area visitor flow transfer probabilities and performs fusion correction on the initial predicted visitor flow.

[0064] Define the correction factor probability weight corr_factor[j] of scenic area j as a random variable, representing the random predicted visitor flow of scenic area j under the combined effect of all external random factors: rand_flow[j] = total_flow × corr_factor[j], where total_flow is the actual total number of people in all valid orders.

[0065] The correction factor probability weights corr_factor[j] for each scenic area are determined as follows: ; Where `max_spot` represents the total number of scenic spots within the hotel area, and `rand_score[j]` can be flexibly defined according to different scenarios. For example, when considering crowd avoidance, it can be adjusted based on the degree of crowding; when considering spatial route convenience, it can be adjusted based on user group trajectory clustering results; when considering weather, the score can be increased or decreased according to the type of scenic spot (indoor / outdoor). The specific calculation method is dynamically configured by the implementer based on available data sources. `tourist_volume[j]` is the average historical visitor volume of scenic spot j during the same period, and it is required that `h1 < 1`.

[0066] `rand_score[j]` is used to quantify the combined impact of external random factors on the probability of visiting scenic spot `j`. Its value is obtained by a weighted sum of multiple computable sub-factors. This application provides a default implementation: rand_score[j] = w1 · f_crowd(j) + w2 · f_history(j) + w3 · f_weather(j) + w4 · f_convenience(j); Where f_crowd(j) = - Init_flow[j] / max_flow, f_crowd(j) represents the negative impact of the current predicted crowding level of scenic spot j, and max_flow is the maximum value of Init_flow among all scenic spots. The larger this value is, the lower the probability that tourists will choose this scenic spot to avoid crowding. f_history(j) = tourist_volume[j] / max_tourist_volume, where tourist_volume[j] is the average daily tourist volume of scenic spot j in the same period of history, and max_tourist_volume is the maximum value of this value among all scenic spots, reflecting the positive guidance of historical popularity on random selection; f_weather(j) is determined based on the scenic area type label. If the scenic area type includes outdoor-dominant labels such as "natural scenery", "theme park", "zoo and botanical garden", and the predicted target date has rain, snow or high temperature weather, then the value is -0.2; otherwise, the value is 0.1. If it is an indoor or semi-indoor type such as "historical culture", "religious temple", "leisure and vacation", then the weather impact value is 0.05. The specific threshold can be dynamically adjusted by the system based on meteorological data. f_convenience(j) = 1 - dist[i][j] / R, which is the normalized spatial distance score, reflecting the random tendency of tourists to choose the scenic spot due to its proximity or convenience.

[0067] Each weight coefficient w1~w4 satisfies w1+w2+w3+w4=1. The system defaults to an equal weight of 0.25. Implementers can fine-tune the corresponding weights according to the reliability of the actual data source (e.g., whether there is real-time weather data), but fine-tuning will not affect the core calculation logic of this scheme.

[0068] Using the above formula, rand_score[j] can be clearly quantified, thus uniquely determining corr_factor[j] and ensuring the feasibility of the passenger flow correction steps. If the data source for a certain sub-factor is missing, the weights corresponding to that factor are reset to zero, and the weights are renormalized among the remaining factors to ensure numerical stability.

[0069] The corrected final flow rate (Final_flow[j]) is: Final_flow[j]=Init_flow[j]*c1+rand_flow[j]*(1-c1); Here, c1 is a balance coefficient (0 ≤ c1 ≤ 1), used to adjust the contribution ratio of the initial predicted passenger flow to the random predicted passenger flow in the final prediction. When c1 approaches 1, the prediction result relies more on the initial predicted passenger flow based on preferences and spatial distance, suitable for stable scenarios with less influence from external random factors; when c1 approaches 0, it is more biased towards passenger flow deviations caused by random factors (such as real-time congestion, weather changes, etc.), suitable for scenarios dominated by sudden random events. In practical applications, c1 can be dynamically set by fitting historical data (e.g., minimizing prediction error) or based on management experience. The default recommended value is 0.6~0.8 to maintain the dominance of initial preferences while retaining the flexibility of random disturbances.

[0070] Prediction results output: The system outputs the final predicted visitor flow Final_flow[j] for each scenic spot on the target date target_date, in "person-times", rounded to one decimal place.

[0071] Understandably, by constructing a user's dwell time window using check-in and check-out times from hotel booking data, and based on the hotel's geographical location and the dwell time window, the potential activity range of users is locked within the dual boundaries of time and space. This achieves a spatial mapping from "hotel check-in data" to "potential scenic spot visitors," breaking through the bottleneck of traditional methods' reliance on scenic spot ticket booking data and significantly extending the lead time for visitor flow prediction. Based on the user identifier and the hotel's geographical location, user preferences are extracted and candidate scenic spots are screened to determine the final set of candidate scenic spots. This achieves structured extraction of user preferences and accurate semantic matching of scenic spot attributes, improving the personalization and rationality of candidate scenic spot selection. The visitor tendency score of each candidate scenic spot in the final set of candidate scenic spots is calculated, providing a comprehensive evaluation basis with a unified metric for subsequent visitor flow probability allocation. Based on the dwell time window, the final set of candidate scenic spots, and the visitor tendency score, the visitor flow of scenic spots on the target date is predicted, obtaining the final predicted visitor flow of each scenic spot on the target date. This allows for the output of predicted visitor flow of each scenic spot before users have booked scenic spot tickets, improving the accuracy and timeliness of scenic spot visitor flow prediction.

[0072] This application has the following beneficial effects: Firstly, by acquiring multi-source data and extracting spatiotemporal windows, based on check-in time, check-out time and hotel geographical location in hotel orders, a user stay time window and spatial neighborhood constraints are constructed. This accurately locks the predicted target group within the dual boundaries of time and space, realizing the spatial mapping from "hotel check-in data" to "potential scenic spot visitors". This breaks through the bottleneck of traditional methods' reliance on scenic spot ticket booking data and significantly extends the lead time for visitor flow prediction.

[0073] Secondly, based on users' historical scenic spot visit records, a user preference vector with a unified semantic scale is constructed through frequency statistics and L1 normalization of the multi-label scenic spot type system. Then, the semantic matching degree between the preference vector and the scenic spot type label vector is calculated by using dot product operation. This transforms the discrete and unstructured historical browsing records into quantifiable and comparable preference feature indicators, realizing the structured extraction of user preferences and accurate semantic matching of scenic spot attributes, thereby improving the personalization and rationality of candidate scenic spot selection.

[0074] Furthermore, by employing a multi-factor weighted fusion method, the three factors of preference semantic matching degree, normalized spatial distance, and global normalized scenic spot popularity are synergistically integrated into a unified visitor tendency score. This overcomes the difficulty of directly integrating the three factors, which belong to different semantic spaces and data scales, into a decision-making process. The visitor tendency score simultaneously reflects the degree of user preference matching, spatial accessibility, and scenic spot popularity, providing a comprehensive evaluation basis with a unified measurement benchmark for subsequent visitor flow probability allocation.

[0075] Finally, based on visitor preference scores, probability normalization allocation is performed within candidate scenic areas, achieving a reasonable initial allocation of visitor flow in scenarios without ticket purchase records. Simultaneously, a random correction factor is introduced to uniformly model unenumerable external random factors as visitor flow shifts between scenic areas. Based on this, a weighted fusion of the initial predicted visitor flow and the randomly predicted visitor flow is performed using a preset balance coefficient. This ensures that the final predicted visitor flow retains the deterministic patterns based on user preferences and spatial constraints while effectively incorporating the impact of unpredictable random disturbances such as real-time congestion and weather changes, thus more closely approximating the mixed probability distribution of actual visitor decisions. The entire method can output the predicted visitor flow for each scenic area before users have booked tickets, providing a longer lead time for public management tasks such as visitor flow control, traffic scheduling, and security deployment.

[0076] Based on the above embodiments, this application also provides a scenic area visitor flow prediction system based on hotel orders and user preferences. Figure 2 As shown, Figure 2 This application provides a schematic diagram of the structure of a scenic area visitor flow prediction system based on hotel orders and user preferences. The system includes: an acquisition module 201, a determination module 202, a calculation module 203, and a prediction module 204. The acquisition module 201 is used to acquire hotel order data from different platforms; wherein, the hotel order data includes user identifier, hotel geographical location, check-in time and check-out time; The determining module 202 is used to determine the user's stay time window at the destination based on the check-in time and the check-out time; wherein, there are multiple users; and based on the user identifier and the hotel's geographical location, user preference extraction and candidate scenic spot filtering are performed to determine the final candidate scenic spot set; The calculation module 203 is used to calculate the visitation preference score for each candidate scenic spot in the final candidate scenic spot set. The prediction module 204 is used to predict the visitor flow of scenic spots on a target date based on the stay time window, the final candidate scenic spot set, and the visitor tendency score, so as to obtain the final predicted visitor flow of each scenic spot on the target date.

[0077] In this embodiment, the determining module 202 is further configured to obtain the user's historical scenic spot visit records based on the user identifier; wherein, the historical scenic spot visit records include scenic spot names and scenic spot type tags; based on the scenic spot names and scenic spot type tags, user preferences are extracted to obtain a user preference vector; based on the hotel's geographical location, the user preference vector, and a spatial radius threshold, candidate scenic spots are screened to determine the final candidate scenic spot set.

[0078] In this embodiment, the determining module 202 is further configured to: calculate the spatial distance between the user's hotel and each scenic spot based on the hotel's geographical location and the geographical coordinates of each scenic spot pre-stored in the scenic spot database; and filter candidate scenic spots in the spatial neighborhood based on the spatial distance between the user's hotel and each scenic spot, as well as the spatial radius threshold, to obtain a set of candidate scenic spots surrounding the user; determine the scenic spot type label vector of each scenic spot in the set of candidate scenic spots surrounding the user; calculate the semantic matching degree of the user's preference with each scenic spot based on the user's preference vector and the scenic spot type label vector of each scenic spot; and filter the scenic spots in the set of candidate scenic spots surrounding the user through the semantic matching degree of the user's preference with each scenic spot and the preference threshold, to determine the final set of candidate scenic spots.

[0079] In this embodiment, the calculation module 203 is further configured to determine the preference semantic matching degree of each candidate scenic spot in the final candidate scenic spot set based on the preference semantic matching degree between the user and each scenic spot; calculate the scenic spot popularity score and the normalized score of spatial distance for each candidate scenic spot in the final candidate scenic spot set; and calculate the access tendency score for each candidate scenic spot in the final candidate scenic spot set based on the scenic spot popularity score, the normalized score of spatial distance, and the preference semantic matching degree.

[0080] In this embodiment, the calculation module 203 is further configured to: determine the spatial distance between each candidate scenic spot in the final candidate scenic spot set and the user's hotel based on the spatial distance between the user's hotel and each scenic spot; calculate the normalized score of the spatial distance of each candidate scenic spot in the final candidate scenic spot set based on the spatial distance between each candidate scenic spot and the user's hotel, and a spatial radius threshold; obtain the historical average daily visitor flow of all scenic spots in the scenic spot database; determine the global minimum and global maximum values ​​based on the historical average daily visitor flow of all scenic spots; and perform normalization processing based on the historical average daily visitor flow of all scenic spots, the global minimum, and the global maximum value to determine the scenic spot popularity score of each candidate scenic spot in the final candidate scenic spot set.

[0081] In this embodiment, the prediction module 204 is further configured to: filter users based on the dwell time window and the target date to determine valid users; determine the final candidate scenic spot set and access tendency score of the valid users; calculate the preference weight of each candidate scenic spot in the final candidate scenic spot set of the valid users based on the final candidate scenic spot set and access tendency score; calculate the initial predicted visitor flow of each candidate scenic spot based on the valid users and the preference weight of each candidate scenic spot; and perform visitor flow correction on the initial predicted visitor flow of each candidate scenic spot to obtain the final predicted visitor flow of each scenic spot on the target date.

[0082] In this embodiment, the prediction module 204 is further configured to obtain the actual total number of people in all valid orders and the historical average visitor flow of each candidate scenic spot during the same period; calculate the correction factor probability weight of each candidate scenic spot based on the historical average visitor flow of each candidate scenic spot and the initial predicted visitor flow of each candidate scenic spot; calculate the random predicted visitor flow of each candidate scenic spot based on the actual total number of people and the correction factor probability weight of each candidate scenic spot; and perform visitor flow correction on the initial predicted visitor flow of each candidate scenic spot based on the random predicted visitor flow of each candidate scenic spot to obtain the final predicted visitor flow of each scenic spot on the target date.

[0083] Based on the above embodiments, this application also provides a scenic area visitor flow prediction device based on hotel orders and user preferences. Figure 3 As shown, Figure 3This is a schematic diagram of a scenic area visitor flow prediction device based on hotel orders and user preferences, provided in an embodiment of this application. The device 3 includes a processor 301 and a memory 302. The memory 302 stores a computer program; the processor 301 retrieves and runs the computer program from the memory to execute the scenic area visitor flow prediction method based on hotel orders and user preferences as described in the above embodiment.

[0084] In the embodiments of this application, the processor 301 described above can be at least one of the following: Application-Specific Integrated Circuit (ASIC), Digital Signal Processor (DSP), Digital Signal Processing Device (DSPD), Programmable Logic Device (PLD), Field Programmable Gate Array (FPGA), Central Processing Unit (CPU), Controller, Microcontroller, and Microprocessor. It is understood that for different devices, the electronic device used to implement the above processor function can also be other types, and the embodiments of this application do not specifically limit it.

[0085] This application provides a computer-readable storage medium storing a computer program for implementing, when executed by a processor, a method for predicting tourist flow in scenic areas based on hotel orders and user preferences as described in any of the above embodiments.

[0086] For example, the program instructions corresponding to the scenic area visitor flow prediction method based on hotel orders and user preferences in this embodiment can be stored on storage media such as optical discs, hard disks, and USB flash drives. When the program instructions corresponding to the scenic area visitor flow prediction method based on hotel orders and user preferences in the storage media are read or executed by an electronic device, the scenic area visitor flow prediction method based on hotel orders and user preferences as described in any of the above embodiments can be realized.

[0087] Furthermore, in the embodiments of this application, the functional modules can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional module.

[0088] If the integrated unit is implemented as a software functional module and is not sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this embodiment, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) or processor to execute all or part of the steps of the method of this embodiment. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0089] It should be understood that the phrases "one embodiment," "an embodiment," or "some embodiments" mentioned throughout the specification mean that a specific feature, structure, or characteristic related to the embodiment is included in at least one embodiment of this application. Therefore, "in one embodiment," "in one embodiment," or "in some embodiments" appearing throughout the specification do not necessarily refer to the same embodiment. Furthermore, these specific features, structures, or characteristics can be combined in any suitable manner in one or more embodiments. It should be understood that in the various embodiments of this application, the sequence numbers of the above-described processes do not imply a sequential order of execution; the execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application. The sequence numbers of the embodiments in this application are merely for descriptive purposes and do not represent the superiority or inferiority of the embodiments. The descriptions of the various embodiments above tend to emphasize the differences between the various embodiments; their similarities or commonalities can be referred to mutually, and for the sake of brevity, these will not be repeated here.

[0090] The modules described above as separate components may or may not be physically separate. The components shown as modules may or may not be physical modules. They may be located in one place or distributed across multiple network units. Some or all of the modules may be selected to achieve the purpose of this embodiment according to actual needs.

[0091] In addition, each functional module in the various embodiments of this application can be integrated into one processing unit, or each module can be a separate unit, or two or more modules can be integrated into one unit; the integrated modules can be implemented in hardware or in the form of hardware plus software functional units.

[0092] Those skilled in the art will understand that all or part of the steps of the above method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it performs the steps of the above method embodiments. The aforementioned storage medium includes various media that can store program code, such as mobile storage devices, read-only memory (ROM), magnetic disks, or optical disks.

[0093] The methods disclosed in the several method embodiments provided in this application can be arbitrarily combined without conflict to obtain new method embodiments.

[0094] The features disclosed in the several product embodiments provided in this application can be arbitrarily combined without conflict to obtain new product embodiments.

[0095] The features disclosed in the several method or device embodiments provided in this application can be arbitrarily combined without conflict to obtain new method or device embodiments.

[0096] The above description is merely an embodiment of this application, but the protection scope of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the protection scope of this application. Therefore, the protection scope of this application should be determined by the protection scope of the claims.

Claims

1. A method for predicting tourist flow in scenic areas based on hotel orders and user preferences, characterized in that, include: Acquire hotel order data from different platforms; wherein, the hotel order data includes user identifier, hotel location, check-in time and check-out time; Based on the check-in time and check-out time, a user's stay time window at the destination is determined; wherein, the number of users is multiple; Based on the user identifier and the hotel's geographical location, user preferences are extracted and candidate scenic spots are filtered to determine the final set of candidate scenic spots; Calculate the visitation preference score for each candidate scenic spot in the final candidate scenic spot set; Based on the dwell time window, the final candidate scenic spot set, and the visitor tendency score, the scenic spot visitor flow is predicted for the target date, and the final predicted visitor flow for each scenic spot on the target date is obtained.

2. The method according to claim 1, characterized in that, The process of extracting user preferences and filtering candidate scenic spots based on the user identifier and the hotel's geographical location to determine the final set of candidate scenic spots includes: Based on the user identifier, the user's historical scenic spot visit records are obtained; wherein, the historical scenic spot visit records include scenic spot name and scenic spot type tag; Based on the scenic area name and the scenic area type label, user preferences are extracted to obtain a user preference vector; Based on the hotel's geographical location, the user preference vector, and the spatial radius threshold, candidate scenic spots are screened to determine the final set of candidate scenic spots.

3. The method according to claim 2, characterized in that, The process of filtering candidate scenic spots based on the hotel's geographical location, the user preference vector, and a spatial radius threshold to determine the final set of candidate scenic spots includes: Based on the hotel's geographical location and the geographical coordinates of each scenic spot pre-stored in the scenic spot database, the spatial distance between the user's hotel and each scenic spot is calculated; and based on the spatial distance between the user's hotel and each scenic spot, as well as the spatial radius threshold, candidate scenic spots in the spatial neighborhood are filtered to obtain a set of candidate scenic spots around the user. Determine the scenic area type label vector for each scenic area in the surrounding candidate scenic area set; Based on the user preference vector and the scenic area type label vector of each scenic area, calculate the semantic matching degree of the user's preference with each scenic area; By using the semantic matching degree and preference threshold between the user and each scenic spot, the scenic spots in the surrounding candidate scenic spot set are filtered to determine the final candidate scenic spot set.

4. The method according to claim 1, characterized in that, The calculation of the visit preference score for each candidate scenic spot in the final candidate scenic spot set includes: Based on the semantic matching degree of the user's preference with each scenic spot, the semantic matching degree of the preference of each candidate scenic spot in the final candidate scenic spot set is determined; Calculate the popularity score and normalized spatial distance score of each candidate scenic spot in the final candidate scenic spot set; Based on the popularity score, the normalized score of spatial distance, and the semantic matching degree of preference for each candidate scenic spot, the visitation tendency score of each candidate scenic spot in the final candidate scenic spot set is calculated.

5. The method according to claim 4, characterized in that, The calculation of the scenic spot popularity score and the normalized score of spatial distance for each candidate scenic spot in the final candidate scenic spot set includes: Based on the spatial distance between the user's hotel and each scenic spot, the spatial distance between each candidate scenic spot in the final candidate scenic spot set and the user's hotel is determined. Based on the spatial distance between each candidate scenic spot and the user's hotel, and the spatial radius threshold, the normalized score of the spatial distance of each candidate scenic spot in the final candidate scenic spot set is calculated. Obtain the historical average daily visitor volume of all scenic spots in the scenic spot database; and determine the global minimum and global maximum values ​​based on the historical average daily visitor volume of all scenic spots. Based on the historical average daily visitor flow of all scenic spots, the global minimum and the global maximum, normalization processing is performed to determine the popularity score of each candidate scenic spot in the final candidate scenic spot set.

6. The method according to claim 1, characterized in that, The process of predicting visitor flow for a target date based on the dwell time window, the final candidate scenic area set, and the visitor preference score, to obtain the final predicted visitor flow for each scenic area on the target date, includes: Users are filtered based on the dwell time window and the target date to determine valid users; Determine the final candidate scenic spot set and access preference score of the effective user; and calculate the preference weight of each candidate scenic spot in the final candidate scenic spot set of the effective user based on the final candidate scenic spot set and access preference score. Based on the effective users and the preference weights of each candidate scenic spot, the initial predicted visitor flow of each candidate scenic spot is calculated. The initial predicted visitor flow for each candidate scenic spot is corrected to obtain the final predicted visitor flow for each scenic spot on the target date.

7. The method according to claim 6, characterized in that, The step of revising the initial predicted visitor flow for each candidate scenic spot to obtain the final predicted visitor flow for each scenic spot on the target date includes: Obtain the actual total number of people with all valid orders and the historical average visitor flow for each candidate scenic spot during the same period; Based on the historical average visitor flow of each candidate scenic spot and the initial predicted visitor flow of each candidate scenic spot, the probability weight of the correction factor for each candidate scenic spot is calculated. Based on the actual total number of people and the probability weights of the correction factors for each candidate scenic spot, the random predicted flow of people for each candidate scenic spot is calculated. Based on the random predicted visitor flow of each candidate scenic spot, the initial predicted visitor flow of each candidate scenic spot is corrected to obtain the final predicted visitor flow of each scenic spot on the target date.

8. A tourist flow prediction system for scenic areas based on hotel orders and user preferences, characterized in that, include: The module comprises an acquisition module, a determination module, a calculation module, and a prediction module; among which, The acquisition module is used to acquire hotel order data from different platforms; wherein, the hotel order data includes user identifier, hotel geographical location, check-in time and check-out time; The determining module is used to determine the user's stay time window at the destination based on the check-in time and the check-out time; wherein, there are multiple users; based on the user identifier and the hotel's geographical location, user preference extraction and candidate scenic spot screening are performed to determine the final candidate scenic spot set; The calculation module is used to calculate the visitation preference score for each candidate scenic spot in the final candidate scenic spot set. The prediction module is used to predict the visitor flow of scenic spots on a target date based on the stay time window, the final candidate scenic spot set, and the visitor tendency score, so as to obtain the final predicted visitor flow of each scenic spot on the target date.

9. A tourist flow prediction device for scenic spots based on hotel orders and user preferences, characterized in that, include: Processor and memory, of which, The memory is used to store computer programs; The processor is configured to call and run the computer program from the memory to perform the method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, It stores executable instructions for causing a processor to execute, thereby implementing the method of any one of claims 1 to 7.

Images