CN111382217A

CN111382217A - Destination recommendation method and device

Info

Publication number: CN111382217A
Application number: CN201811636707.6A
Authority: CN
Inventors: 卢凯敏
Original assignee: Beijing Qihoo Technology Co Ltd
Current assignee: Beijing Qihoo Technology Co Ltd
Priority date: 2018-12-29
Filing date: 2018-12-29
Publication date: 2020-07-07

Abstract

The invention discloses a destination recommendation method, a destination recommendation device, electronic equipment and a computer-readable storage medium. The method comprises the following steps: acquiring historical driving track data of a user; constructing a destination recommendation model by taking the acquired historical driving track data as training sample data; when receiving an inquiry request sent by the user, determining a destination recommended to the user according to the current position information and the current time information of the user by using the constructed destination recommendation model and displaying the destination recommended to the user. In the technical scheme, the destination recommendation model is constructed by taking the historical driving track data of the user as reference, the travel habit of the user is met, the recommended destination determined by the destination recommendation model is more consistent with the travel requirement of the user, the step that the user needs manual input can be avoided, the method is particularly suitable for being applied to intelligent hardware or equipment of travel, and the use experience of the user is improved.

Description

Destination recommendation method and device

Technical Field

The invention relates to the technical field of computers, in particular to a destination recommendation method, a destination recommendation device, electronic equipment and a computer-readable storage medium.

Background

The existing navigation equipment or trip application programs are favored by users because of bringing great convenience to the users. In order to implement the navigation function, it is necessary to determine a destination to which a user is to arrive, and then make a travel decision according to the determined destination. In the prior art, in a mode of determining a destination, a user inputs the destination through a manual or voice line, however, the mode has a large operation cost, especially, navigation equipment such as a car recorder navigation hardware is inconvenient for the user to directly input manually, even if the probability of input errors due to environment, hardware size and the like is greatly increased after input, not only is a bad use experience brought to the user, but also the operability of the equipment is reduced.

Disclosure of Invention

In view of the above, the present invention has been made to provide a destination recommendation method, apparatus, electronic device, and computer-readable storage medium that overcome or at least partially solve the above problems.

According to an aspect of the present invention, there is provided a destination recommendation method, wherein the method includes:

acquiring historical driving track data of a user;

constructing a destination recommendation model by taking the acquired historical driving track data as training sample data;

when receiving an inquiry request sent by the user, determining a destination recommended to the user according to the current position information and the current time information of the user by using the constructed destination recommendation model and displaying the destination recommended to the user.

Optionally, the acquiring historical driving track data of the user includes:

the device ID, timestamp data, and location information of the user are obtained.

Optionally, the constructing a destination recommendation model by using the acquired historical driving track data as training sample data includes:

determining one or more candidate destination information according to the acquired historical driving track data;

and calculating the score of each candidate destination according to the determined candidate destination information, and taking the preset number of candidate destinations with the highest scores as the destinations to be recommended of the destination recommendation model.

Optionally, the determining one or more candidate destination information according to the acquired historical driving track data includes:

determining one or more starting and ending point pairs according to the acquired historical driving track data;

performing hierarchical clustering on the determined one or more starting and ending point pairs to obtain one or more candidate cluster sets;

and merging the obtained candidate clusters, and determining candidate destination information corresponding to each candidate cluster.

Optionally, the determining one or more starting and ending point pairs according to the acquired historical driving track data includes:

determining one or more sub-driving track data corresponding to the user according to the acquired historical driving track data;

and obtaining one or more starting and ending point pairs according to the determined one or more sub-driving track data.

Optionally, the hierarchically clustering the determined one or more starting and ending point pairs, and the determining one or more candidate cluster sets includes:

performing first-layer clustering on the determined one or more starting and ending point pairs by using a hierarchical clustering training model to obtain one or more candidate regions;

performing second-layer clustering on the obtained one or more candidate areas, and determining a candidate sub-area with the maximum staying position density in each candidate area;

and taking each staying position in each candidate sub-region as an element of a candidate cluster set corresponding to the candidate sub-region to obtain a candidate cluster set corresponding to each candidate sub-region.

Optionally, the merging the obtained candidate clusters, and determining candidate destination information corresponding to each candidate cluster includes:

and calculating the center point of each candidate cluster set, and taking the calculated center point of each candidate cluster set as a candidate destination corresponding to the candidate cluster set.

Optionally, the calculating a score of each candidate destination according to the determined candidate destination information includes:

calculating a score of each candidate destination according to the determined candidate destination information by using a score function; wherein the scoring function comprises:

wherein S is_iIs the score of the candidate destination, β is a time decay parameter, t_nowIs the model training date; t is t_lastIs the date the user last stayed at the candidate destination; p (x) is a parameter of the sigmoid function,

loc_timeis the number of points of the candidate destination in the corresponding candidate cluster.

Optionally, the method further comprises:

training destination sample data labeled with category labels to obtain a destination label classification model;

acquiring specified characteristics of a preset number of candidate destinations with highest scores;

and determining the classification label of each candidate destination by using the destination label classification model according to the acquired specified characteristics of each candidate destination.

Optionally, the specified characteristics include one or more of:

the number of the driving tracks;

the probability of an early peak occurring;

probability of late peak occurrence;

probability of occurrence of a weekday;

probability of weekend occurrence;

location point POI category attributes.

Optionally, the method further comprises:

and adding reverse geocoding information to the acquired preset number of candidate destinations with the highest scores.

Optionally, the determining, by using the constructed destination recommendation model according to the current location information and the current time information of the user, a destination recommended to the user includes:

and calculating the score of each recommendation destination in the destination recommendation model by using the constructed destination recommendation model according to the current position information and the current time information of the user, and taking the destination to be recommended with the highest score as the destination recommended to the user.

Optionally, the calculating, by using the constructed destination recommendation model according to the current location information and the current time information of the user, a score of each recommendation destination in the destination recommendation model includes:

calculating the score of each candidate destination by using a score function according to the current position information and the current time information of the user; wherein the scoring function comprises:

wherein S is_i' is the current score of the candidate destination, β is a time decay parameter, t_now' is current time information; t is t_lastIs the date the user last stayed at the candidate destination; p (x) is a parameter of the sigmoid function,

Optionally, the method further comprises:

storing the constructed destination recommendation model to a specified database, and providing a corresponding application programming API (application programming interface);

when receiving the query request sent by the user, determining the destination recommended to the user by using the constructed destination recommendation model according to the current position information and the current time information of the user comprises:

and when receiving a query request sent by the user, calling the API interface, and determining a destination recommended to the user by using the constructed destination recommendation model according to the current position information and the current time information of the user.

Optionally, the method further comprises:

and carrying out noise reduction processing on the acquired historical driving track data.

Optionally, the method further comprises:

and smoothing the historical driving track data subjected to the denoising treatment.

Optionally, the method further comprises:

receiving a path inquiry instruction for a specified destination input by the user or receiving a selection instruction for a recommended destination by the user;

and according to the received instruction, determining one or more pieces of path information from the current position of the user to the destination and displaying the path information to the user.

According to another aspect of the present invention, there is provided a destination recommendation apparatus, wherein the apparatus includes:

the device comprises an acquisition unit, a display unit and a control unit, wherein the acquisition unit is suitable for acquiring historical driving track data of a user;

the model construction unit is suitable for constructing a destination recommendation model by taking the acquired historical driving track data as training sample data;

and the destination determining unit is suitable for determining a destination recommended to the user and displaying the destination to the user by utilizing the constructed destination recommendation model according to the current position information and the current time information of the user when receiving the query request sent by the user.

Alternatively,

the acquisition unit is suitable for acquiring the equipment ID, the time stamp data and the position information of the user.

Alternatively,

the model building unit is suitable for determining one or more candidate destination information according to the acquired historical driving track data; and calculating the score of each candidate destination according to the determined candidate destination information, and taking the preset number of candidate destinations with the highest scores as the destinations to be recommended of the destination recommendation model.

Alternatively,

the model building unit is suitable for determining one or more starting and ending point pairs according to the acquired historical driving track data; performing hierarchical clustering on the determined one or more starting and ending point pairs to obtain one or more candidate cluster sets; and merging the obtained candidate clusters, and determining candidate destination information corresponding to each candidate cluster.

Alternatively,

the model building unit is suitable for determining one or more sub-driving track data corresponding to the user according to the acquired historical driving track data; and obtaining one or more starting and ending point pairs according to the determined one or more sub-driving track data.

Alternatively,

the model building unit is suitable for performing first-layer clustering on the determined one or more starting and ending point pairs by utilizing a hierarchical clustering training model to obtain one or more candidate regions; performing second-layer clustering on the obtained one or more candidate areas, and determining a candidate sub-area with the maximum staying position density in each candidate area; and taking each staying position in each candidate sub-region as an element of a candidate cluster set corresponding to the candidate sub-region to obtain a candidate cluster set corresponding to each candidate sub-region.

Alternatively,

the model building unit is suitable for calculating the center point of each candidate cluster set, and taking the calculated center point of each candidate cluster set as a candidate destination corresponding to the candidate cluster set.

Alternatively,

the model construction unit is suitable for calculating scores of all candidate destinations according to the determined candidate destination information by using a score function; wherein the scoring function comprises:

Alternatively,

the model construction unit is suitable for training destination sample data marked with category labels to obtain a destination label classification model; acquiring specified characteristics of a preset number of candidate destinations with highest scores; and determining the classification label of each candidate destination by using the destination label classification model according to the acquired specified characteristics of each candidate destination.

Optionally, the specified characteristics include one or more of:

the number of the driving tracks;

the probability of an early peak occurring;

probability of late peak occurrence;

probability of occurrence of a weekday;

probability of weekend occurrence;

location point POI category attributes.

Alternatively,

the model construction unit is suitable for adding reverse geocoding information to the obtained preset number of candidate destinations with the highest scores.

Alternatively,

and the destination determining unit is suitable for calculating the score of each recommendation destination in the destination recommendation model by using the constructed destination recommendation model according to the current position information and the current time information of the user, and taking the destination to be recommended with the highest score as the destination recommended to the user.

Alternatively,

the destination determining unit is suitable for calculating the score of each candidate destination by using a score function according to the current position information and the current time information of the user; wherein the scoring function comprises:

wherein S is_i' is the current score of the candidate destination, β is a time decay parameter, t_now' is the current time letterInformation; t is t_lastIs the date the user last stayed at the candidate destination; p (x) is a parameter of the sigmoid function,

Optionally, the apparatus further comprises:

the interface providing unit is suitable for storing the constructed destination recommendation model to a specified database and providing a corresponding application programming API interface;

and the destination determining unit is suitable for calling the API when receiving the query request sent by the user, and determining the destination recommended to the user by using the constructed destination recommendation model according to the current position information and the current time information of the user.

Optionally, the apparatus further comprises:

and the preprocessing unit is suitable for performing noise reduction processing on the acquired historical driving track data.

Alternatively,

the preprocessing unit is suitable for smoothing the historical driving track data subjected to denoising processing.

Optionally, the apparatus further comprises:

the route determining unit is suitable for receiving a route inquiry instruction which is input by the user and used for specifying a destination or receiving a selection instruction of the user and used for recommending the destination; and according to the received instruction, determining one or more pieces of path information from the current position of the user to the destination and displaying the path information to the user.

According to still another aspect of the present invention, there is provided an electronic apparatus, wherein the electronic apparatus includes:

a processor; and the number of the first and second groups,

a memory arranged to store computer executable instructions that, when executed, cause the processor to perform a method according to the foregoing.

According to yet another aspect of the present invention, there is provided a computer readable storage medium, wherein the computer readable storage medium stores one or more programs which, when executed by a processor, implement the aforementioned method.

According to the technical scheme of the invention, historical driving track data of a user is obtained; constructing a destination recommendation model by taking the acquired historical driving track data as training sample data; when receiving an inquiry request sent by the user, determining a destination recommended to the user according to the current position information and the current time information of the user by using the constructed destination recommendation model and displaying the destination recommended to the user. In the technical scheme, the destination recommendation model is constructed by taking the historical driving track data of the user as reference, the travel habit of the user is met, the recommended destination determined by the destination recommendation model is more consistent with the travel requirement of the user, the step that the user needs manual input can be avoided, the method is particularly suitable for being applied to intelligent hardware or equipment of travel, and the use experience of the user is improved.

The foregoing description is only an overview of the technical solutions of the present invention, and the embodiments of the present invention are described below in order to make the technical means of the present invention more clearly understood and to make the above and other objects, features, and advantages of the present invention more clearly understandable.

Drawings

Various other advantages and benefits will become apparent to those of ordinary skill in the art upon reading the following detailed description of the preferred embodiments. The drawings are only for purposes of illustrating the preferred embodiments and are not to be construed as limiting the invention. Also, like reference numerals are used to refer to like parts throughout the drawings. In the drawings:

FIG. 1 shows a flow diagram of a destination recommendation method according to one embodiment of the invention;

FIG. 2 shows a flow diagram of a destination recommendation method according to another embodiment of the invention;

fig. 3 shows a schematic configuration diagram of a destination recommendation device according to an embodiment of the present invention;

FIG. 4 shows a schematic structural diagram of an electronic device according to one embodiment of the invention;

fig. 5 shows a schematic structural diagram of a computer-readable storage medium according to an embodiment of the present invention.

Detailed Description

Exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be embodied in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.

Fig. 1 shows a flow diagram of a destination recommendation method according to an embodiment of the invention. As shown in fig. 1, the method includes:

and step S110, acquiring historical driving track data of the user.

And step S120, constructing a destination recommendation model by taking the acquired historical driving track data as training sample data.

In this embodiment, the historical trajectory data of the user is accumulated continuously, so that the destination recommendation model trained and constructed according to the historical trajectory data also changes along with the accumulated change of the historical trajectory data, and thus the constructed destination recommendation model can be guaranteed to change along with the change of the travel habits of the user in real time, and the requirements of the user are met better.

Step S130, when receiving the query request sent by the user, determining a destination recommended to the user according to the current location information and the current time information of the user by using the constructed destination recommendation model, and displaying the destination recommended to the user.

Here, the query request sent by the user may be understood as a login request when the user logs in a corresponding application program, or the query request is sent when the navigation hardware is started.

For a scheme of destination recommendation, in the prior art, a recommendation list returned based on a destination recently input by a user is provided for the user to select, but in this way, the history input of the user is used as a reference, and some history inputs may be destinations that the user rarely goes, so that it cannot be ensured that a destination meeting the needs of the user exists in the recommendation list every time, which causes inaccurate destination recommendation, and reduces the user experience. For example, the user has recently input "the great wall of octal" as the destination, but the destination is only input as the recent playing requirement of the user, and if the "great wall of octal" in the recommendation list exists all the time thereafter, the requirement of the user is not met, the recommendation resource is wasted, and the use experience of the user is reduced. Or the user inputs the 'Shanghai station' as a destination recently, the user only intends to check the distance or path information between the position where the user is and the 'Shanghai station', and the user does not really want to go to the 'Shanghai station', and then the 'Shanghai station' is listed in the recommendation list, so that the user requirements are not met.

In the embodiment, the travel habit of the user can be determined after training based on the historical driving track data of the user instead of the historical input of the user, the destination frequently visited by the user is obtained, the destination recommendation model is built according to the travel habit, the optimal destination is recommended according to the current position and time of the user, the user requirement can be better met, even if the historical driving track data contains the destination which the user does not frequently go (such as the octadina great wall in the above example), the historical driving track data is only one of the historical driving track data, and the repeated data is not much or does not exist, and the building of the destination recommendation model cannot be influenced.

Therefore, according to the embodiment, the destination recommendation model is constructed by taking the historical driving track data of the user as the reference, the travel habit of the user is met, the recommended destination is determined more accurately by using the destination recommendation model, the steps that the user needs to input manually can be avoided, the method is particularly suitable for being applied to intelligent hardware or equipment of travel, the recommended destination can be guaranteed to better meet the travel requirement of the user, and the use experience of the user is improved.

In an embodiment of the present invention, the acquiring historical driving path data of the user in step S110 of the method shown in fig. 1 includes: the device ID, timestamp data, and location information of the user are obtained.

The location information includes DPS information, time stamp information includes time information at a location, and information such as a time of stay at the location and a time of travel between different locations can be determined by the time stamp information. For example, the historical driving trace data obtained is day 18: 00, position 1; the next day 8: 00, position 1, 8: 02, position 2. It is determined that the stay at position 1 was 14 hours and 2 minutes was driven from position 1 to position 2.

In an embodiment of the present invention, the constructing the destination recommendation model by using the acquired historical driving path data as training sample data in step S120 of the method shown in fig. 1 includes: determining one or more candidate destination information according to the acquired historical driving track data; and calculating the score of each candidate destination according to the determined candidate destination information, and taking the preset number of candidate destinations with the highest scores as the destinations to be recommended of the destination recommendation model.

In this embodiment, one or more candidate destinations are determined from the historical driving path data, and in consideration of the fact that the determined candidate destinations are frequently used by users and are not frequently used by users, in this embodiment, it is further necessary to calculate ranking scores for determining each candidate destination, and select a plurality of candidate destinations with the highest scores from the candidate destinations as the destinations to be recommended in the destination recommendation model, that is, in response to an inquiry request from a user, only the best destination is selected again from the selected destinations to be recommended by using the destination recommendation model, and other candidate destinations do not need to be considered.

For example, 10 candidate destinations are determined, and after the ranking score of each candidate destination is calculated, 5 candidate destinations are sequentially selected from top to bottom as destinations to be recommended.

Specifically, the determining one or more candidate destination information according to the acquired historical driving track data includes: determining one or more starting and ending point pairs according to the acquired historical driving track data; performing hierarchical clustering on the determined one or more starting and ending point pairs to obtain one or more candidate cluster sets; and merging the obtained candidate clusters, and determining candidate destination information corresponding to each candidate cluster.

The obtained historical trajectory data includes one or more trajectories of the user, and each trajectory includes a start point and an end point, i.e., an origin and a destination, so that one or more start-end point pairs can be determined from the historical trajectory data, where the start-end point pair is a start-end point pair, and can be expressed as (start point, end point), and the start point and the end point of each start-end point pair are corresponding. For example, three starting and ending point pairs are determined according to the historical driving track data, including (starting point 1, ending point 1), (starting point 2, ending point 2), and (starting point 3, ending point 3).

After hierarchical clustering is carried out on the starting and ending point pairs, each starting and ending point pair corresponds to a candidate cluster set, each candidate cluster set comprises a plurality of position information obtained after hierarchical clustering, and then candidate destination information corresponding to each candidate cluster set is determined according to the position information.

Preferably, the determining one or more starting and ending point pairs according to the acquired historical driving track data includes: determining one or more sub-driving track data corresponding to the user according to the acquired historical driving track data; and obtaining one or more starting and ending point pairs according to the determined one or more sub-driving track data.

Considering that the acquired historical driving track data includes the position information and the timestamp information, the historical driving track data needs to be divided according to the historical driving track data to determine one or more corresponding sub-driving track data, wherein the sub-driving track data are single-day maximum interval sub-driving track data. Each piece of sub-trajectory data includes start point information, end point information, driving duration, parking duration, and positioning times corresponding to the sub-trajectory. The starting point and the end point of each sub-track form a starting and end point pair.

For example, the trajectory data includes: 8: 00, position 1, 8: 02, position 2, 8: 30-17: 00, position 3, 17:02, position 4, 17: 30-24: 00, position 1, it may be determined that parking at position 3 is 8 and a half hours, it may be determined that one sub-trajectory is from position 1-position 2-position 3, and after 17:30 of the day, parking at position 1 is 6 and a half hours, it may be determined that another sub-trajectory is from position 3-position 4-position 5. The starting point information position 1, the end point information position 3, the running time 30 minutes, the end point parking time 8 and half hours and the positioning times of the first sub-driving track can be obtained through the driving track data; the starting point information position 3 and the end point information position 1 of the second sub-driving track, the driving time length of 30 minutes, the end point parking time length of 6 and half hours and the positioning times.

Preferably, the above hierarchically clustering the determined one or more starting and ending point pairs, and the determining one or more candidate cluster sets includes: performing first-layer clustering on the determined one or more starting and ending point pairs by using a hierarchical clustering training model to obtain one or more candidate regions; performing second-layer clustering on the obtained one or more candidate areas, and determining a candidate sub-area with the maximum staying position density in each candidate area; and taking each staying position in each candidate sub-region as an element of a candidate cluster set corresponding to the candidate sub-region to obtain a candidate cluster set corresponding to each candidate sub-region.

The hierarchical clustering model comprises clustering algorithms such as MeanShift and DBSCAN. In this embodiment, two-layer clustering is performed on the determined one or more starting and ending point pairs, a first-layer clustering adopts a DBSCAN clustering algorithm, one or more candidate regions can be determined according to the one or more starting and ending point pairs, and candidate destinations can be obtained from the candidate regions. And then performing second-layer clustering on the obtained one or more candidate areas, and determining a candidate sub-area with the highest parking position density from the candidate areas by adopting a MeanShift clustering algorithm, wherein the candidate sub-area is considered that the user does not stay at the same position every time when arriving at a destination, for example, in a home, the user may park a vehicle in a parking lot of the cell, and if the parking lot has no parking space, the vehicle may also be parked outside the cell. Therefore, in order to accurately obtain the candidate destination, the candidate sub-region with the highest density of the stay positions in each candidate region needs to be determined through a clustering algorithm, it is known that the determined candidate sub-region also includes a plurality of stay positions, and each stay position is used as an element of the candidate cluster set corresponding to the candidate sub-region to obtain the candidate cluster set corresponding to each candidate sub-region.

Preferably, the merging the obtained candidate clusters, and determining the candidate destination information corresponding to each candidate cluster includes: and calculating the center point of each candidate cluster set, and taking the calculated center point of each candidate cluster set as a candidate destination corresponding to the candidate cluster set.

Because the candidate cluster set includes a plurality of stay positions, in order to obtain an accurate candidate destination, the center points of all stay positions in each candidate cluster set are calculated to serve as candidate destinations corresponding to the candidate cluster set, and accordingly, the area radius, the positioning times, the positioning days and the time distribution of each candidate cluster set need to be determined so as to grasp the information of the candidate cluster set integrally. Parameter support is provided for the determination of the destination to be recommended.

Preferably, the calculating the score of each candidate destination according to the determined candidate destination information includes: calculating the score of each candidate destination according to the determined candidate destination information by using a ranking rank score function; wherein the scoring function comprises:

wherein S is_iIs the score of the candidate destination, β is a time decay parameter, typically β ═ 0.98, t_nowIs the model training date; t is t_lastIs the date the user last stayed at the candidate destination, t_now-t_lastIs the difference between the model training date and the date the user last appeared at the candidate destination; p (x) is a parameter of the sigmoid function,

For example, 10 candidate destinations are determined, the ranking score is calculated for each candidate destination using the above-mentioned score function, and the 10 candidate destinations are numbered 1-10 in order, S₁Calculating the ranking score of the candidate destination with the label of 1; s₂The ranking scores of the candidate destinations corresponding to the reference numeral 2 are calculated, and by analogy, the ranking scores of 10 candidate destinations are obtained.

As described above, the historical trajectory data of the user is accumulated continuously, and therefore, the destination recommendation model trained and constructed according to the historical trajectory data also changes with the accumulated change of the historical trajectory data, that is, the destination to be recommended in the destination recommendation model also changes in real time, that is, the score of each candidate destination is calculated differently if one or more pieces of candidate destination information determined according to the obtained historical trajectory data are different. For example, the candidate destinations determined on the first day are position 1, position 2, position 3, and the calculated score position 1> position 2> position 3; the candidate destinations determined the next day are again position 1, position 2, position 3, the calculated score is position 2> position 1> position 3; alternatively, the candidate destinations determined the next day are again position 1, position 4, position 5, and the calculated score is position 4> position 5> position 1.

In one embodiment of the present invention, the method shown in fig. 1 further comprises: training destination sample data labeled with category labels to obtain a destination label classification model; acquiring specified characteristics of a preset number of candidate destinations with highest scores; and determining the classification label of each candidate destination by using a destination label classification model according to the acquired specified characteristics of each candidate destination.

After obtaining the preset number of destinations to be recommended, in order to perform more accurate recommendation, in this embodiment, classification labels are added to the preset number of destinations to be recommended, such as homes, companies, and the like, specifically, labeling is performed by using a destination label classification model, and in order to obtain the destination label classification model, it is necessary to obtain a plurality of destinations labeled with class labels as training samples, and train the model. Specifically, training may be performed according to the specified features of the destinations to which the classification labels are labeled, and then the same features of the destinations to be recommended also need to be obtained when classifying and labeling the destinations to be recommended.

In particular, the specified features described above include one or more of: the number of the driving tracks; the probability of an early peak occurring; probability of late peak occurrence; probability of occurrence of a weekday; probability of weekend occurrence; location point POI category attributes.

The location point POI (point of interest) category attribute can also be used as one of the references of the classification label, for example, if the POI type belongs to a business area, the probability of the company is greater than the probability of the family.

For example, the probability of occurrence of the destination to be recommended on a weekday is lower than the probability of occurrence on a weekend, and the probability of occurrence in an idle time is greater than the probability of occurrence of a peak in the morning and evening, then the destination to be recommended is labeled as home. Of course, when the destination label classification model is used, the features can be directly input into the model, and then the classification label of the destination to be recommended can be directly acquired.

When the optimal recommended destination is determined according to the query request of the user, the classification label of the destination to be recommended can be used as a reference to ensure the accuracy of the determined recommended destination, for example, at the time of next shift, the classification label can be recommended to the user as the destination to be recommended of the home, and the non-classification label is the destination to be recommended of the company.

In one embodiment of the present invention, the method shown in fig. 1 further comprises: and adding reverse geocoding information to the acquired preset number of candidate destinations with the highest scores.

Geocoding refers to the process of representing a detailed address of a place name in geographic coordinates (e.g., latitude and longitude). The process of mapping the address information into geographic coordinates is called geocoding; the process of converting the geographic coordinates to address information is referred to as reverse geocoding. In the present embodiment, the obtained candidate destination is a geographical coordinate, and therefore, it is necessary to add inverse geocoding information to the candidate destination in order to convert the candidate destination into address information.

In one embodiment of the present invention, the determining, by using the constructed destination recommendation model, the destination recommended to the user according to the current location information and the current time information of the user in step S130 shown in fig. 1 includes: and calculating the score of each recommendation destination in the destination recommendation model by using the constructed destination recommendation model according to the current position information and the current time information of the user, and taking the destination to be recommended with the highest score as the destination recommended to the user.

Because the destination recommendation model includes a preset number of destinations to be recommended, when an inquiry request of a user is received, an optimal destination is selected from the preset number of destinations to be recommended and recommended to the user, therefore, the ranking scores of the recommended destinations need to be respectively calculated according to the current position information and the current time information of the user, and the destination to be recommended with the highest score is taken as the destination recommended to the user. In the same destination recommendation model, if the current location information and the current time information of the user are different, the determined recommended destinations are also different. For example, the destination recommendation model includes a destination 1, a destination 2, and a destination 3, and if the user is currently at the location 1 and at the time 1, the determined destination recommended to the user is the destination 1; if the user is at the position 2 and at the time 2, the determined destination recommended to the user is the destination 2.

Specifically, the calculating the score of each recommended destination in the destination recommendation model according to the current position information and the current time information of the user by using the constructed destination recommendation model includes: calculating the score of each candidate destination by using a score function according to the current position information and the current time information of the user; wherein the scoring function comprises:

wherein S is_i' is the current score of the candidate destination, β is a time decay parameter, typically β ═ 0.98, t_now' is current time information; t is t_lastIs the date the user last stayed at the candidate destination; p (x) is a parameter of the sigmoid function,

In the present embodiment, the score function used is the same as the score function for calculating the candidate destinations of the destination recommendation model, except that t is the same here_now' is the current time information (the time when the user sent the query request) and is not the date of model training. Other parameters may be referred to in the above description.

In one embodiment of the present invention, the method shown in fig. 1 further comprises: and saving the constructed destination recommendation model to a specified database, and providing a corresponding application programming API interface.

When receiving the query request sent by the user in step S130, determining, by using the constructed destination recommendation model, a destination recommended to the user according to the current location information and the current time information of the user includes: and when receiving a query request sent by the user, calling an API (application programming interface), and determining a destination recommended to the user by using the constructed destination recommendation model according to the current position information and the current time information of the user.

The designated database here includes a redis database. And calling the destination recommendation model through an API interface.

In one embodiment of the present invention, the method shown in fig. 1 further comprises: and carrying out noise reduction processing on the acquired historical driving track data.

In this embodiment, after the historical driving track data is acquired, in order to ensure the accuracy of the data, noise reduction processing needs to be performed on the acquired historical driving track data to remove noise points.

Further, the method shown in fig. 1 further comprises: and smoothing the historical driving track data subjected to the denoising treatment.

In this embodiment, a kalman filter algorithm is used to perform track smoothing on the noise-reduced historical trajectory data.

In one embodiment of the present invention, the method shown in fig. 1 further comprises: receiving a path inquiry instruction for a specified destination input by the user or receiving a selection instruction for a recommended destination by the user; and according to the received instruction, determining one or more pieces of path information from the current position of the user to the destination and displaying the path information to the user.

In this embodiment, considering that the destination recommended to the user by the technical solution is not required by the user, the user may manually input or input voice to specify the destination, and query a path to the specified destination. Or, the destination recommended to the user by the technical scheme is required by the user, and the user can select the destination and inquire a path to the specified destination. Therefore, in the embodiment, the path information of the destination input by the user or selected by the user is determined and displayed to the user, so that the user can go to the destination according to the determined path information, and the use experience of the user is further improved.

Fig. 2 shows a flow diagram of a destination recommendation method according to another embodiment of the invention. As shown in fig. 2, in step S210, a client APP records travel track GPS data of a user, including a device ID, a timestamp, and GPS longitude and latitude information; step S220, the client uploads the data to a cloud server in real time, and the accumulated data is used as a training data source; constructing a destination recommendation model according to the training data source: step S230, performing track noise reduction on a plurality of historical tracks of the same user, removing noise points, and step S240, performing track smoothing on the historical track data after drying is removed by using a Kalman filtering algorithm; step S250, segmenting the smoothed track data to obtain a plurality of maximum sub-tracks, and extracting start _ location and end _ location starting point pairs and end _ pairs of each sub-track; step S260, two-layer clustering is performed on a plurality of loc _ calls starting and ending points by using MeanShift, DBSCAN and the like; step S270, calculating characteristics such as the central points and time distribution of various clustered clusters, the area radius, the positioning times, the positioning days and the like to obtain a plurality of candidate destinations; step S280, calculating rank scores of all candidate destinations; step S290, selecting top5 candidate destinations from all the candidate destinations to obtain a destination recommendation model; step S201, calling an inverse geocoding service, and adding inverse geocoding information to a candidate destination top5 in a destination recommendation model; step S202, predicting the category label of the top5 candidate destination by using a category label model; step S203, storing the model data into redis and providing a destination recommendation service API interface; step S204, when a user opens the client device, acquiring the device ID, the GPS information and the timestamp information, and sending an http request to a server; step S205, the server calls a destination recommendation service API; step S206, acquiring a current destination recommendation list of the user through a destination recommendation service; and step S207, the client performs voice interaction determination with the user based on the optimal destination result, and determines path information after the determination.

In a specific example, the client sends a real-time request, and calculates that the score of the destination to be recommended, namely 'golden garden', is 0.7834 highest, the score of the destination to be recommended, namely 'ylange station', is 0.1745, and the client can preferentially recommend the 'golden garden' to the user for making a final travel decision. Further verification is carried out, the fact that the user is not located nearby the residence currently and is determined to be 6:30 in the afternoon from the real-time request, and the server judges that the most possible travel intention of the user is a 'returning' strategy by combining historical GPS track behaviors of the user, so that the ranking score calculated by the destination recommendation model is reasonable.

Fig. 3 shows a schematic configuration diagram of a destination recommendation device according to an embodiment of the present invention. As shown in fig. 3, the destination recommendation apparatus 300 includes:

the obtaining unit 310 is adapted to obtain historical driving trace data of a user.

The model building unit 320 is adapted to build the destination recommendation model by using the acquired historical driving track data as training sample data.

And the destination determining unit 330 is adapted to, when receiving the query request sent by the user, determine a destination recommended to the user according to the current location information and the current time information of the user by using the constructed destination recommendation model and show the destination recommended to the user.

In an embodiment of the present invention, the obtaining unit 310 shown in fig. 3 is adapted to obtain the device ID, the timestamp data and the location information of the user.

In an embodiment of the present invention, the model building unit 320 shown in fig. 3 is adapted to determine one or more candidate destination information according to the acquired historical driving track data; and calculating the score of each candidate destination according to the determined candidate destination information, and taking the preset number of candidate destinations with the highest scores as the destinations to be recommended of the destination recommendation model.

Specifically, the model building unit 320 is adapted to determine one or more starting and ending point pairs according to the acquired historical driving track data; performing hierarchical clustering on the determined one or more starting and ending point pairs to obtain one or more candidate cluster sets; and merging the obtained candidate clusters, and determining candidate destination information corresponding to each candidate cluster.

Preferably, the model building unit 320 is adapted to determine one or more sub-trajectory data corresponding to the user according to the acquired historical trajectory data; and obtaining one or more starting and ending point pairs according to the determined one or more sub-driving track data.

Preferably, the model building unit 320 is adapted to perform first-level clustering on the determined one or more starting and ending point pairs by using a hierarchical clustering training model to obtain one or more candidate regions; performing second-layer clustering on the obtained one or more candidate areas, and determining a candidate sub-area with the maximum staying position density in each candidate area; and taking each staying position in each candidate sub-region as an element of a candidate cluster set corresponding to the candidate sub-region to obtain a candidate cluster set corresponding to each candidate sub-region.

Preferably, the model building unit 320 is adapted to calculate a central point of each candidate cluster, and use the calculated central point of each candidate cluster as a candidate destination corresponding to the candidate cluster.

Preferably, the model building unit 320 is adapted to calculate the score of each candidate destination according to the determined candidate destination information by using the ranking rank score function; wherein the scoring function comprises:

In an embodiment of the present invention, the model building unit 320 shown in fig. 3 is adapted to train destination sample data labeled with a category label to obtain a destination label classification model; acquiring specified characteristics of a preset number of candidate destinations with highest scores; and determining the classification label of each candidate destination by using a destination label classification model according to the acquired specified characteristics of each candidate destination.

27. The apparatus of claim 26, wherein the specified characteristics include one or more of: the number of the driving tracks; the probability of an early peak occurring; probability of late peak occurrence; probability of occurrence of a weekday; probability of weekend occurrence; location point POI category attributes.

In an embodiment of the present invention, the model building unit 320 shown in fig. 3 is adapted to add inverse geocoding information to the obtained preset number of candidate destinations with the highest scores.

In an embodiment of the present invention, the destination determining unit 330 shown in fig. 3 is adapted to calculate a score of each recommended destination in the destination recommendation model according to the current location information and the current time information of the user by using the constructed destination recommendation model, and take the destination to be recommended with the highest score as the destination recommended to the user.

Specifically, the destination determining unit 330 is adapted to calculate the score of each candidate destination by using a score function according to the current location information and the current time information of the user; wherein the scoring function comprises:

In one embodiment of the present invention, the apparatus shown in fig. 3 further comprises:

and the interface providing unit is suitable for saving the constructed destination recommendation model to a specified database and providing a corresponding application programming API interface.

And the destination determining unit 330 is adapted to, when receiving the query request sent by the user, call the API interface, and determine a destination recommended to the user by using the constructed destination recommendation model according to the current location information and the current time information of the user.

In one embodiment of the present invention, the apparatus shown in fig. 3 further comprises: and the preprocessing unit is suitable for performing noise reduction processing on the acquired historical driving track data.

Further, the preprocessing unit is suitable for smoothing the historical driving track data after denoising processing.

In summary, according to the technical scheme of the invention, historical driving track data of a user is acquired; constructing a destination recommendation model by taking the acquired historical driving track data as training sample data; when receiving an inquiry request sent by the user, determining a destination recommended to the user according to the current position information and the current time information of the user by using the constructed destination recommendation model and displaying the destination recommended to the user. In the technical scheme, the destination recommendation model is constructed by taking the historical driving track data of the user as reference, the travel habit of the user is met, the recommended destination determined by the destination recommendation model is more consistent with the travel requirement of the user, the step that the user needs manual input can be avoided, the method is particularly suitable for being applied to intelligent hardware or equipment of travel, and the use experience of the user is improved.

It should be noted that:

the algorithms and displays presented herein are not inherently related to any particular computer, virtual machine, or other apparatus. Various general purpose devices may be used with the teachings herein. The required structure for constructing such a device will be apparent from the description above. Moreover, the present invention is not directed to any particular programming language. It is appreciated that a variety of programming languages may be used to implement the teachings of the present invention as described herein, and any descriptions of specific languages are provided above to disclose the best mode of the invention.

In the description provided herein, numerous specific details are set forth. It is understood, however, that embodiments of the invention may be practiced without these specific details. In some instances, well-known methods, structures and techniques have not been shown in detail in order not to obscure an understanding of this description.

Similarly, it should be appreciated that in the foregoing description of exemplary embodiments of the invention, various features of the invention are sometimes grouped together in a single embodiment, figure, or description thereof for the purpose of streamlining the disclosure and aiding in the understanding of one or more of the various inventive aspects. However, the disclosed method should not be interpreted as reflecting an intention that: that the invention as claimed requires more features than are expressly recited in each claim. Rather, as the following claims reflect, inventive aspects lie in less than all features of a single foregoing disclosed embodiment. Thus, the claims following the detailed description are hereby expressly incorporated into this detailed description, with each claim standing on its own as a separate embodiment of this invention.

Those skilled in the art will appreciate that the modules in the device in an embodiment may be adaptively changed and disposed in one or more devices different from the embodiment. The modules or units or components of the embodiments may be combined into one module or unit or component, and furthermore they may be divided into a plurality of sub-modules or sub-units or sub-components. All of the features disclosed in this specification (including any accompanying claims, abstract and drawings), and all of the processes or elements of any method or apparatus so disclosed, may be combined in any combination, except combinations where at least some of such features and/or processes or elements are mutually exclusive. Each feature disclosed in this specification (including any accompanying claims, abstract and drawings) may be replaced by alternative features serving the same, equivalent or similar purpose, unless expressly stated otherwise.

Furthermore, those skilled in the art will appreciate that while some embodiments described herein include some features included in other embodiments, rather than other features, combinations of features of different embodiments are meant to be within the scope of the invention and form different embodiments. For example, in the following claims, any of the claimed embodiments may be used in any combination.

The various component embodiments of the invention may be implemented in hardware, or in software modules running on one or more processors, or in a combination thereof. It will be appreciated by those skilled in the art that a microprocessor or Digital Signal Processor (DSP) may be used in practice to implement some or all of the functions of some or all of the components of the destination recommendation apparatus, electronic device and computer readable storage medium according to embodiments of the present invention. The present invention may also be embodied as apparatus or device programs (e.g., computer programs and computer program products) for performing a portion or all of the methods described herein. Such programs implementing the present invention may be stored on computer-readable media or may be in the form of one or more signals. Such a signal may be downloaded from an internet website or provided on a carrier signal or in any other form.

For example, fig. 4 shows a schematic structural diagram of an electronic device according to an embodiment of the invention. The electronic device 400 conventionally includes a processor 410 and a memory 420 arranged to store computer-executable instructions (program code). The memory 420 may be an electronic memory such as a flash memory, an EEPROM (electrically erasable programmable read only memory), an EPROM, a hard disk, or a ROM. The memory 420 has a memory space 430 that stores program code 440 for performing any of the method steps shown in fig. 1 or fig. 2 and in various embodiments. For example, the memory space 430 for the program code may include respective program codes 440 for implementing respective steps in the above method. The program code can be read from or written to one or more computer program products. These computer program products comprise a program code carrier such as a hard disk, a Compact Disc (CD), a memory card or a floppy disk. Such a computer program product is generally a computer-readable storage medium 500 such as described in fig. 5. The computer-readable storage medium 500 may have memory segments, memory spaces, etc. arranged similarly to the memory 420 in the electronic device of fig. 4. The program code may be compressed, for example, in a suitable form. In general, the memory unit stores program code 510 for performing the steps of the method according to the invention, i.e. program code readable by a processor such as 410, which when run by an electronic device causes the electronic device to perform the steps of the method described above.

It should be noted that the above-mentioned embodiments illustrate rather than limit the invention, and that those skilled in the art will be able to design alternative embodiments without departing from the scope of the appended claims. In the claims, any reference signs placed between parentheses shall not be construed as limiting the claim. The word "comprising" does not exclude the presence of elements or steps not listed in a claim. The word "a" or "an" preceding an element does not exclude the presence of a plurality of such elements. The invention may be implemented by means of hardware comprising several distinct elements, and by means of a suitably programmed computer. In the unit claims enumerating several means, several of these means may be embodied by one and the same item of hardware. The usage of the words first, second and third, etcetera do not indicate any ordering. These words may be interpreted as names.

Claims

1. A destination recommendation method, wherein the method comprises:

acquiring historical driving track data of a user;

2. The method of claim 1, wherein the obtaining historical driving trajectory data of the user comprises:

3. The method according to any one of claims 1-2, wherein the constructing a destination recommendation model by using the acquired historical driving track data as training sample data comprises:

4. The method of any one of claims 1-3, wherein the determining one or more candidate destination information from the obtained historical trajectory data comprises:

5. A destination recommendation apparatus, wherein the apparatus comprises:

6. The apparatus of claim 5, wherein,

7. The apparatus of any one of claims 5-6,

8. The apparatus of any one of claims 5-7,

9. An electronic device, wherein the electronic device comprises:

a processor; and the number of the first and second groups,

a memory arranged to store computer executable instructions that, when executed, cause the processor to perform the method of any one of claims 1 to 4.

10. A computer readable storage medium, wherein the computer readable storage medium stores one or more programs which, when executed by a processor, implement the method of any of claims 1-4.