Floor recommendation method, electronic device, and computer-readable storage medium

Abstract

The present application discloses a floor recommendation method, electronic device, and computer-readable storage medium. The floor recommendation method includes: obtaining a first vector, a second vector, and a third vector corresponding to a floor; generating the first vector based on the area type of the floor, generating the second vector based on the elevator usage data corresponding to the current time period, and generating the third vector based on the floor number after encoding; optimizing the prediction model using the first vector, the second vector, and the third vector to obtain an optimized prediction model; obtaining a fourth vector of a preset dimension based on the first vector, the second vector, and the recommendation model, and determining the floor recommended for the elevator to stop in the current time period based on the fourth vector; optimizing the recommendation model using the first vector, the second vector, the fourth vector, and the prediction model to obtain an optimized recommendation model, and returning to the initial step. The above scheme can enhance the adaptability to different floors and improve the accuracy of floor recommendations.

Description

Technical Field

[0001] The present application relates to the field of data mining technology, and in particular to a floor recommendation method, an electronic device, and a computer-readable storage medium. Background Art

[0002] With the progress of urbanization, more and more buildings are using elevators. However, in actual use, elevators often have long waiting times. Existing technologies typically do not process the floors where elevators are idle, or simply stop at a preset fixed floor when idle. Due to the lack of in-depth data mining of elevator usage time periods and usage data, elevators are unable to adaptively adjust the length of time they stay during idle periods based on the floor or time period. As a result, elevators are not highly adaptable to different floors. Therefore, improving adaptability to different floors and improving the accuracy of floor recommendations has become an urgent issue. Summary of the Invention

[0003] The main technical problem solved by this application is to provide a floor recommendation method, electronic device and computer-readable storage medium, which can enhance the adaptability to different floors and improve the accuracy of floor recommendation.

[0004] To solve the above technical problems, the first aspect of the present application provides a floor recommendation method, including: obtaining a first vector, a second vector and a third vector corresponding to the floor; the first vector is generated based on the area type of the floor, the second vector is generated based on the elevator usage data corresponding to the current time period, and the third vector is generated after encoding the floor number of the floor; the prediction model is optimized using the first vector, the second vector and the third vector to obtain an optimized prediction model; based on the first vector, the second vector and the recommendation model, a fourth vector of a preset dimension is obtained, and based on the fourth vector, the floor recommended for the elevator to stop in the current time period is determined; the recommendation model is optimized using the first vector, the second vector, the fourth vector and the prediction model to obtain the optimized recommendation model, and the step of obtaining the first vector, the second vector and the third vector corresponding to the floor is returned.

[0005] To solve the above technical problems, the second aspect of the present application provides an electronic device, which includes: a memory and a processor coupled to each other, wherein the memory stores program data, and the processor calls the program data to execute the method described in the first aspect above.

[0006] In order to solve the above technical problems, the third aspect of the present application provides a computer storage medium on which program data is stored. When the program data is executed by a processor, the method described in the first aspect is implemented.

[0007] The above scheme generates a first vector based on the area type of the floor, and generates a second vector based on the elevator usage data corresponding to the current time period, wherein the first vector is related to space and the second vector is related to time. A third vector is generated after encoding the floor number based on the floor. The first vector, the second vector and the third vector are used to optimize the prediction model to improve the data mining ability of the prediction model. Based on the first vector related to space, the second vector related to time and the recommendation model, a fourth vector of a preset dimension is obtained. Based on the fourth vector, the floor recommended for the elevator to stop in the current time period can be determined. The first vector, the second vector and the fourth vector and the prediction model are used to optimize the recommendation model to improve the data mining ability of the recommendation model. The prediction model and the recommendation model are continuously optimized in a continuous cycle, thereby enhancing the adaptability to different floors and improving the accuracy of floor recommendations. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For those skilled in the art, other drawings can be obtained based on these drawings without inventive efforts. Among them:

[0009] Figure 1 This is a flow chart of an implementation method of the floor recommendation method of the present application;

[0010] Figure 2 This is a flow chart of another embodiment of the floor recommendation method of the present application;

[0011] Figure 3 This is a schematic structural diagram of an embodiment of the electronic device of the present application;

[0012] Figure 4 It is a structural diagram of an embodiment of the computer storage medium of the present application. DETAILED DESCRIPTION

[0013] The following will be combined with the accompanying drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of them. Based on the embodiments of this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0014] The terms "system" and "network" are often used interchangeably in this document. The term "and / or" is simply a description of an association between related objects, indicating that three possible relationships exist. For example, "A and / or B" can mean: A exists alone, A and B exist simultaneously, or B exists alone. Furthermore, the character " / " generally indicates that the related objects are in an "or" relationship. Furthermore, "multiple" in this document means two or more than two.

[0015] See also Figure 1 , Figure 1 This is a flow chart of an embodiment of a floor recommendation method of the present application, which includes:

[0016] S101: Obtain a first vector, a second vector, and a third vector corresponding to a floor, wherein the first vector is generated based on the area type of the floor, the second vector is generated based on elevator usage data corresponding to the current time period, and the third vector is generated based on the floor number encoding of the floor.

[0017] Specifically, the area types of different areas on the floor are counted, and a first vector corresponding to the floor is generated based on the area types on the floor, wherein the first vector is related to the spatial characteristics corresponding to the floor. When someone outside the elevator presses the elevator button, the time when the elevator is pressed is recorded as elevator usage data. A second vector is generated based on the elevator usage data, wherein the second vector is related to the time characteristics corresponding to the floor, and the floor number corresponding to the elevator when the elevator is pressed is recorded. The floor number of the floor is encoded to generate a third vector.

[0018] In one application method, area types are divided into five types based on the use of the area, including residential area, office area, teaching area, shopping area and medical area. A five-dimensional first vector is generated based on the use of different areas on the floor, wherein the first vector is related to the area type on the floor and the first vector is obtained after normalization processing. Therefore, the first vector is related to the spatial characteristics of the floor.

[0019] Furthermore, the preset cycle is divided into multiple time periods. When someone presses the elevator button outside the elevator in different time periods, the time when the elevator is pressed and the floor where the elevator is located are recorded. When someone presses the elevator button outside the elevator in the current time period, corresponding vector weights are set for the current time period and the two adjacent time periods before and after the current time period, and corresponding vector weights are set for other time periods other than the current time period and the two adjacent time periods. The vector weights corresponding to all time periods are integrated and normalized to obtain a second vector. The second vector is related to the time characteristics of using the elevator on a floor.

[0020] Further, the floors are uniquely hot encoded based on the floor number of the floor to obtain a third vector, wherein the third vector is related to the floor number corresponding to the floor.

[0021] S102: Optimizing the prediction model using the first vector, the second vector, and the third vector to obtain an optimized prediction model.

[0022] Specifically, the first vector, the second vector, and the third vector are connected, and the connected first vector, the second vector, and the third vector are input into the prediction model so that the prediction model outputs a prediction result, wherein the prediction result is the total number of floors to which the elevator moves after someone presses the elevator button outside the elevator when the elevator stops at a certain floor during idle time.

[0023] In one application, an elevator is designated to stop at a preset floor during idle time in the current time period. The total number of floors the elevator has traveled during the current time period is pre-calculated as a statistical result. The concatenated first, second, and third vectors are input into a prediction model. The output layer of the prediction model is a one-dimensional output, and a predicted value is obtained as a prediction result. Based on the error between the predicted result and the statistical result, the parameters in the prediction model are optimized to obtain an optimized prediction model.

[0024] In one application scenario, the prediction model is in an initial state and is optimized to enable the prediction model to have prediction capabilities.

[0025] In another application scenario, the prediction model already has prediction capabilities, and after optimization, its data mining capabilities are continuously improved.

[0026] S103: Based on the first vector, the second vector and the recommendation model, a fourth vector of a preset dimension is obtained, and based on the fourth vector, a floor at which the elevator is recommended to stop in the current time period is determined.

[0027] Specifically, the first vector and the second vector are connected to input into the recommendation model, so that the recommendation model outputs a fourth vector of a preset dimension, wherein the preset dimension is related to the total number of floors of the building where the floor is located.

[0028] In one application, the total number of floors L of the building where the floor is located is obtained in advance, the output layer of the recommendation model is set to L neurons, the first vector and the second vector are connected and input into the recommendation model, and the recommendation model outputs an L-dimensional fourth vector based on the first vector related to spatial features and the second vector related to temporal features.

[0029] Furthermore, an activation function is used to activate the fourth vector of the preset dimension, and a maximum value is extracted from the activated fourth vector as the floor recommended for the elevator to stop in the current time period.

[0030] S104: Optimize the recommendation model using the first vector, the second vector, the fourth vector, and the prediction model to obtain an optimized recommendation model.

[0031] Specifically, the first vector, the second vector and the fourth vector are connected, and the connected first vector, the second vector and the fourth vector are input into the prediction model to obtain a new prediction result corresponding to the prediction model, wherein the fourth vector and the third vector are both related to the floors where the elevator is recommended to stop, and the prediction model outputs a prediction result, which is the total number of floors that the elevator moves in the current time period based on the latest recommended stop floors.

[0032] Furthermore, the parameters of the recommendation model are optimized based on unsupervised learning to minimize the total number of floors the elevator travels during the current time period by setting the most recently recommended floors as the elevator's idle stop locations. This optimizes the recommendation model. The most recently recommended floors are used to replace the previously recommended floors, and the process returns to the step of obtaining the first, second, and third vectors corresponding to the floors.

[0033] It can be understood that when traversing to any time period in the preset cycle, the corresponding time period will be used as the current time period, and when the preset cycle is traversed, it will enter the next cycle of the preset cycle, so that the prediction model and recommendation model can be optimized and updated in real time, and recommendations can be made for the elevator's stop floors in different time periods to improve the accuracy of floor recommendations.

[0034] The above scheme generates a first vector based on the area type of the floor, and generates a second vector based on the elevator usage data corresponding to the current time period, wherein the first vector is related to space and the second vector is related to time. A third vector is generated after encoding the floor number based on the floor. The first vector, the second vector and the third vector are used to optimize the prediction model to improve the data mining ability of the prediction model. Based on the first vector related to space, the second vector related to time and the recommendation model, a fourth vector of a preset dimension is obtained. Based on the fourth vector, the floor recommended for the elevator to stop in the current time period can be determined. The first vector, the second vector and the fourth vector and the prediction model are used to optimize the recommendation model to improve the data mining ability of the recommendation model. The prediction model and the recommendation model are continuously optimized in a continuous cycle, thereby enhancing the adaptability to different floors and improving the accuracy of floor recommendations.

[0035] See also Figure 2 , Figure 2 1 is a flow chart of another embodiment of the floor recommendation method of the present application, which includes:

[0036] S201: Obtain the first vector, the second vector, and the third vector corresponding to the floor.

[0037] Specifically, the first vector is generated based on the area type of the floor, the second vector is generated based on the elevator usage data corresponding to the floor in the current time period, and the third vector is generated based on the floor number encoding of the floor.

[0038] In one application, the step of obtaining the first vector corresponding to the floor includes generating a first initial vector based on an area type on the floor, and normalizing the first initial vector to obtain the first vector. The area type includes at least one of a residential area, an office area, a teaching area, a shopping area, and a medical area.

[0039] Specifically, the areas within each floor of the building and the area types corresponding to the areas are counted, and a corresponding initial weight is set for each area type. When a certain area type is not included in a floor, the initial weight of the corresponding area type is set to 0. The initial weights corresponding to all area types are arranged in a preset order to generate a first initial vector. Data related to the area type on the floor is generated by the initial weight. The area type corresponds to the spatial characteristics on the floor, and the first initial vector is normalized to obtain the first vector.

[0040] In a specific application scenario, the area types corresponding to all areas in the building corresponding to the floor are counted. When there are multiple areas of the same area type, the initial weight of the corresponding area type is set according to the number of areas. When there are three office areas, the initial weight corresponding to the office area is 3. When there is no medical area, the initial weight corresponding to the medical area is 0. The initial weight setting method for other area types is the same and will not be repeated here. The initial weights corresponding to all area types are arranged in a preset order to obtain a first initial vector, and then the first initial vector is normalized to obtain the first vector.

[0041] In another specific application scenario, the area types corresponding to all areas in the building corresponding to the floors are counted, a corresponding first weight is set for each floor, and a corresponding second weight is set for each area type, wherein the first weight is related to the height of the floor, and the second weight is related to the priority of the area type. When the area types include residential areas, office areas, teaching areas, shopping areas, and medical areas, their corresponding priorities can be arranged from large to small to be medical areas, teaching areas, office areas, residential areas, and shopping areas. The initial weights of the corresponding area types are set according to the number of areas, and the adjusted weights are obtained based on the first weights corresponding to the floors and the second weights of the area types. The adjusted weights corresponding to all area types are arranged in a preset order to obtain a first initial vector, and the first initial vector is then normalized to obtain a first vector.

[0042] In one application, the step of obtaining a second vector corresponding to a floor includes: dividing a preset period into a plurality of time periods, obtaining elevator usage data within a current time period; generating a second initial vector based on the elevator usage data, and normalizing the second initial vector to obtain a second vector.

[0043] Specifically, the preset period is divided into multiple time periods, and the usage time of the elevator in the current time period is counted. Excluding discrete random data, the time when users use the elevator is usually regular. Based on the regular data with higher probability, when someone presses the elevator button outside the elevator in the current time period, the corresponding time is counted as the usage time. A second initial vector is generated based on the usage time of the elevator, so that the second initial vector is correlated with the time feature corresponding to the floor, and then the second initial vector is normalized to obtain the second vector, so as to achieve the unity of the first vector and the second vector.

[0044] In one application scenario, the elevator usage data includes the usage time of the elevator. The step of generating a second initial vector based on the elevator usage data includes: obtaining the vector weight of the current time period and the vector weights of two adjacent time periods before and after the current time period based on the usage time of the elevator and the length of the time period; setting the vector weights corresponding to other time periods other than the current time period and the two adjacent time periods to 0 to obtain the second initial vector.

[0045] Specifically, the preset cycle is divided into multiple time periods based on a preset time length, the usage time of the elevator in the current time period is obtained, and two adjacent time periods before the current time period and after the current time period and adjacent to the current time period are extracted. Vector weights are generated for the current time period and the two adjacent time periods based on the usage time of the elevator and the time length of the time period. The vector weights corresponding to other time periods other than the current time period and the two adjacent time periods are set to 0, so that the usage time of the elevator becomes regular data for subsequent mining, and the vector weights are arranged in the order of the time periods to obtain a second initial vector.

[0046] In a specific application scenario, the preset period is 24 hours and the time length is 30 minutes. The preset period is divided into 48 time periods based on the preset time length. When someone presses the elevator button outside the elevator during the current time period, the time the elevator button is pressed is used as the elevator usage time during the current time period. Corresponding vector weights are generated for the current time period and the two adjacent time periods. The vector weights corresponding to other time periods outside the current time period and the two adjacent time periods are set to 0, thereby generating a 48-dimensional second initial vector. The above process is expressed by the following formula:

[0047]

[0048] Among them, Ti is the vector weight, τ is the time length, t m-1 is the median of the adjacent time periods before the current time period, t m+1 is the median of the adjacent time period after the current time period, t m is the median of the current time period, t is the usage time of the elevator, and the median of each time period is the time in the middle of the corresponding time period.

[0049] In one application, the floor numbers corresponding to the floors are uniquely encoded to obtain a third vector, which is related to the floor numbers corresponding to the floors where the elevator stops during idle time in the current time period.

[0050] S202: Input the first vector into the feature extraction model to obtain an updated first vector, and input the second vector into the feature extraction model to obtain an updated second vector.

[0051] Specifically, the first vector is input into multiple fully connected layers in the feature extraction model to obtain an updated first vector, and the second vector is input into multiple one-dimensional convolution structures in the feature extraction model to obtain an updated second vector, so that the first vector and the second vector can mine deeper features after passing through the feature extraction model.

[0052] In one application, the first vector is N-dimensional The second vector is M-dimensional The feature vector After several layers of one-dimensional convolutional structures, M ′ ×c-dimensional matrix, M ′ ×c-dimensional matrices are connected into M ′ ×c-dimensional vector The feature vector After several layers of fully connected networks, we get N ′ ×c-dimensional vector Among them, the first vector after update and the second vector It has a deeper correlation with the spatial and temporal characteristics in the current time period, thereby obtaining the correlation with the environmental characteristics of the elevator in the current time period.

[0053] S203: Connect the updated first vector, the updated second vector, and the third vector to obtain a fifth vector, and input the fifth vector into the prediction model to obtain a first floor motion prediction value.

[0054] Specifically, the elevator usage data includes the average floor movement of elevators during the current time period. The average floor movement of elevators is pre-collected after the elevators report. The updated first vector and the updated second vector are concatenated, and then concatenated with the third vector to obtain a fifth vector. The fifth vector is input into the prediction model, which outputs a predicted value for the first floor movement based on environmental characteristics and floor-related characteristics.

[0055] In one application, the updated first vector is connected and the updated second vector Get the feature vector Among them, the eigenvector The dimension is (N′+M′)×c, and the feature vector With the third vector Connect to get the fifth vector The fifth vector The prediction model is input to obtain a one-dimensional output r, which represents the average value of the elevator movement floors predicted by the prediction model and is recorded as the first floor movement prediction value.

[0056] S204: Optimizing parameters of the feature extraction model and the prediction model based on the difference between the first floor motion prediction value and the floor motion average value to obtain optimized feature extraction model and prediction model.

[0057] Specifically, the difference between the first floor motion prediction value and the floor motion average value is obtained, and the parameters of the feature extraction model and the prediction model are optimized based on the difference, so as to obtain the optimized feature extraction model and prediction model.

[0058] In one application, the difference between the first floor motion prediction value and the floor motion average value is minimized to adjust the parameters in the feature extraction model and the prediction model, thereby improving the prediction model's data mining capabilities and improving the accuracy of the prediction. The above process is expressed as follows:

[0059]

[0060] Where r is the predicted value of the first floor motion and s is the average value of the floor motion.

[0061] S205: Connect the updated first vector and the updated second vector to obtain a sixth vector, and input the sixth vector into the recommendation model to obtain a fourth vector of a preset dimension.

[0062] Specifically, the preset dimension is related to the total number of floors of the building where the floor is located, the updated first vector and the updated second vector are connected to obtain a sixth vector, and the sixth vector is input into the recommendation model to obtain a fourth vector of the preset dimension.

[0063] In one application, the updated first vector is connected and the updated second vector Get the sixth vector The sixth vector Input the recommendation model so that the recommendation model is based on environmental features, that is, the recommendation model outputs a fourth vector of preset dimensions based on spatial and temporal features Among them, when the total number of floors in the building where the floor is located is L, the output layer of the recommendation model is set to L neurons, and the recommendation model outputs the fourth vector of L dimension

[0064] S206: Activate the fourth vector using an activation function to obtain the floor number with the maximum probability value corresponding to the fourth vector as the floor recommended for the elevator to stop in the current time period.

[0065] Specifically, the fourth vector is activated using the softmax function to obtain the floor number with the highest probability as the floor recommended for the elevator to stop in the current time period, so as to update the floors recommended for the elevator to stop in the current time period during this cycle and improve the accuracy of the recommendation.

[0066] S207: Connect the updated first vector, the updated second vector, and the fourth vector to obtain a seventh vector, and input the seventh vector into the optimized prediction model to obtain a second floor motion prediction value.

[0067] Specifically, the updated first vector, the updated second vector and the fourth vector are connected to obtain the seventh vector, wherein the fourth vector is also related to the number of floors. At this time, the fourth vector is related to the latest recommended floor to stay. The seventh vector is input into the optimized prediction model to obtain the second floor motion prediction value.

[0068] In one application, the updated first vector is connected and the updated second vector and the fourth vector Get the seventh vector The seventh vector The data is input to the optimized prediction model, which outputs a second floor motion prediction value based on the feature vector associated with the latest recommended floor.

[0069] S208: Optimize the parameters of the recommendation model based on the second floor motion prediction value to obtain an optimized recommendation model.

[0070] Specifically, based on the second floor motion prediction value, a loss function corresponding to the recommendation model is generated, and the loss is minimized to adjust the parameters in the recommendation model to obtain an optimized recommendation model, thereby improving the recommendation model's data mining capabilities and improving the accuracy of recommendations. The above process is expressed as follows:

[0071]

[0072] Where r′ is the motion prediction value of the second floor.

[0073] Furthermore, the last recommended floor is replaced by the latest recommended floor, thereby returning to the step of obtaining the first vector, the second vector, and the third vector corresponding to the floor.

[0074] In this embodiment, a recommendation model is used to update the floors recommended for the elevator to stop based on a first vector related to spatial features and a second vector related to temporal features, and the recommendation model and the prediction model are optimized during real-time learning to improve the accuracy of floor recommendations in different time periods in different application scenarios, thereby reducing the waiting time for users when using the elevator.

[0075] See also Figure 3 , Figure 3 This is a structural diagram of an embodiment of an electronic device of the present application. The electronic device 30 includes a memory 301 and a processor 302 coupled to each other, wherein the memory 301 stores program data (not shown), and the processor 302 calls the program data to implement the floor recommendation method in any of the above embodiments. For an explanation of the relevant content, please refer to the detailed description of the above method embodiments, which will not be repeated here.

[0076] See also Figure 4 , Figure 4 This is a structural diagram of an embodiment of the computer storage medium of the present application. The computer storage medium 40 stores program data 400. When the program data 400 is executed by the processor, the floor recommendation method in any of the above embodiments is implemented. For an explanation of the relevant content, please refer to the detailed description of the above method embodiments, which will not be repeated here.

[0077] It should be noted that the units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of this embodiment.

[0078] In addition, the functional units in the various embodiments of the present application may be integrated into a single processing unit, or each unit may exist physically separately, or two or more units may be integrated into a single unit. The aforementioned integrated units may be implemented in the form of hardware or software functional units.

[0079] If the integrated unit is implemented in the form of a software functional unit and 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 the present application, 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, which is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) or a processor to execute all or part of the steps of each embodiment method of the present application. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0080] The above description is only an implementation method of the present application and does not limit the patent scope of the present application. Any equivalent structure or equivalent process transformation made using the contents of the description and drawings of this application, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present application.

Claims

1. A floor recommendation method, characterized in that: The method comprises: Obtain a first vector, a second vector, and a third vector corresponding to a floor; the first vector is generated based on the area type of the floor, the second vector is generated based on elevator usage data corresponding to the current time period, and the third vector is generated based on the floor number encoding of the floor; Optimizing the prediction model using the first vector, the second vector, and the third vector to obtain an optimized prediction model; Based on the first vector, the second vector and the recommendation model, a fourth vector of a preset dimension is obtained, and based on the fourth vector, a floor at which the elevator is recommended to stop in the current time period is determined; The recommendation model is optimized using the first vector, the second vector, the fourth vector, and the prediction model to obtain the optimized recommendation model, and the step of obtaining the first vector, the second vector, and the third vector corresponding to the floor is returned; wherein the prediction result of the recommendation model includes the floors at which the elevator is recommended to stop, and the prediction result of the prediction model includes the total number of floors the elevator moves in the current time period based on the latest recommended stop floors.

2. The floor recommendation method according to claim 1, characterized in that: The step of obtaining a first vector corresponding to a floor includes: generating a first initial vector based on an area type on the floor, and normalizing the first initial vector to obtain the first vector; wherein the area type includes at least one of a residential area, an office area, a teaching area, a shopping area, and a medical area; The step of obtaining a second vector corresponding to a floor includes: dividing a preset period into a plurality of time periods, obtaining elevator usage data within a current time period; generating a second initial vector based on the elevator usage data, and normalizing the second initial vector to obtain the second vector.

3. The floor recommendation method according to claim 2, characterized in that: The elevator usage data includes the usage time of the elevator, and the step of generating the second initial vector based on the elevator usage data includes: Based on the use time of the elevator and the length of the time period, obtaining a vector weight of the current time period and vector weights of two adjacent time periods before and after the current time period; The vector weights corresponding to the current time period and other time periods other than the two adjacent time periods are set to 0 to obtain the second initial vector.

4. The floor recommendation method according to claim 1, characterized in that: Before the step of optimizing the prediction model using the first vector, the second vector, and the third vector to obtain an optimized prediction model, the method further includes: The first vector is input into a feature extraction model to obtain an updated first vector, and the second vector is input into the feature extraction model to obtain an updated second vector.

5. The floor recommendation method according to claim 4, characterized in that: The elevator usage data includes the floor movement average value of the elevator movement in the current time period; The step of optimizing the prediction model using the first vector, the second vector, and the third vector to obtain an optimized prediction model includes: connecting the updated first vector, the updated second vector, and the third vector to obtain a fifth vector, and inputting the fifth vector into the prediction model to obtain a first floor motion prediction value; Based on the difference between the first floor motion prediction value and the floor motion average value, the parameters of the feature extraction model and the prediction model are optimized to obtain the optimized feature extraction model and the prediction model.

6. The floor recommendation method according to claim 4, characterized in that: The step of obtaining a fourth vector of a preset dimension based on the first vector, the second vector, and the recommendation model includes: The updated first vector and the updated second vector are connected to obtain a sixth vector, and the sixth vector is input into the recommendation model to obtain a fourth vector of a preset dimension; wherein the preset dimension is related to the total number of floors of the building where the floor is located.

7. The floor recommendation method according to claim 6, characterized in that: The step of optimizing the recommendation model by using the first vector, the second vector, the fourth vector and the prediction model to obtain the optimized recommendation model includes: Connecting the updated first vector, the updated second vector, and the fourth vector to obtain a seventh vector, and inputting the seventh vector into the optimized prediction model to obtain a second floor motion prediction value; Optimize the parameters of the recommendation model based on the second floor motion prediction value to obtain the optimized recommendation model.

8. The floor recommendation method according to claim 1, characterized in that: The step of determining the floor at which the elevator is recommended to stop within the current time period based on the fourth vector includes: The fourth vector is activated by using an activation function to obtain the floor number with the maximum probability value corresponding to the fourth vector as the floor recommended for the elevator to stop in the current time period.

9. An electronic device, characterized in that: include: A memory and a processor coupled to each other, wherein the memory stores program data, and the processor calls the program data to execute the method according to any one of claims 1 to 8.

10. A computer-readable storage medium having program data stored thereon, characterized in that: When the program data is executed by a processor, the method according to any one of claims 1 to 8 is implemented.

Images