Pickup volume warning method, device, terminal equipment and storage medium

Abstract

The present invention discloses a pickup volume warning method, which is applied to smart express lockers. The method includes: obtaining and saving weather information, and when a user's account information is detected, determining the information of the packages to be picked up based on the account information; predicting the user's load capacity based on the account information and the weather information; and outputting a warning prompt message when it is detected that the number of packages to be picked up corresponding to the information of the packages to be picked up exceeds the user's load capacity. The present invention also discloses a pickup volume warning device, a terminal device, and a storage medium. The present invention can combine weather and user factors to predict the user's single pickup volume, thereby improving the accuracy of the prediction of the user's pickup volume. When the packages to be picked up exceed the user's load capacity, a warning prompt message is output to prompt the user to reduce the pickup volume, thereby preventing the user from picking up packages that exceed their own load capacity at one time, affecting their normal walking, and thus improving the user experience.

Description

Technical Field

[0001] The present invention relates to the field of machine learning technology, and in particular to a pickup volume warning method, apparatus, terminal device, and storage medium. Background Art

[0002] With the rapid development of internet technology, the large volume of packages generated by online shopping has provided a promising application foundation for smart express lockers. When it's inconvenient for users to sign for a package, they can temporarily store it in a locker and retrieve it at their leisure time upon receiving a pickup notification. The emergence of smart lockers not only solves the last-mile delivery problem but also greatly facilitates package pickup. However, due to limitations in the installation sites of smart lockers, the pickup location is often some distance from the user's home or office. Users may use umbrellas on sunny or rainy days, which can affect the number of packages they can pick up in a single trip. Furthermore, users may not accurately estimate the size or weight of their packages when picking them up, and even indoors, they may not be able to accurately predict outdoor weather conditions, which can make walking difficult, thus impacting the user experience. For privacy reasons, we cannot fully monitor the user registration and pickup process and collect information such as their location. However, the number of items a user will pick up in a single trip is affected by numerous factors, including gender, age, transportation used, and weather. Therefore, the only way to predict a user's pickup volume is through model prediction. Existing prediction models are not closely aligned with business characteristics and are prone to overfitting. Therefore, they cannot accurately estimate a user's pickup volume. This can lead to users being unable to walk after picking up items because the amount they pick up exceeds their capacity. Summary of the Invention

[0003] The main purpose of the present invention is to provide a pickup volume warning method, device, terminal device and storage medium, aiming to solve the technical problem of inaccurate prediction of user pickup volume in the existing technology.

[0004] In addition, to achieve the above-mentioned purpose, the present invention also provides a pickup volume early warning method, which is applied to a smart express locker. The pickup volume early warning method includes the following steps:

[0005] Acquire and save weather information, and when the user's account information is detected, determine the information of the package to be picked up based on the account information;

[0006] predicting the user's load capacity based on the account information and the weather information;

[0007] When it is detected that the number of packages to be picked up corresponding to the package information to be picked up exceeds the load capacity of the user, an early warning prompt message is output.

[0008] Optionally, the step of predicting the user's load capacity based on the account information and the weather information includes:

[0009] inputting the account information and the weather information into a target prediction model, and extracting target feature information from the account information and the weather information using the target prediction model;

[0010] The user's load capacity is predicted based on the target feature information, wherein the target prediction model is obtained by iteratively training a preset basic prediction model using a sample data set, and the sample data set is generated based on each user's account information, weather information at the time of pickup, and feedback information after pickup.

[0011] Optionally, before the step of inputting the account information and the weather information into the target prediction model, the method further includes:

[0012] Obtaining feedback information from a first user group after picking up a package, where the first user group is composed of users who generate feedback information after picking up a package;

[0013] A sample data set is generated according to the feedback information, and a preset basic prediction model is iteratively trained using the sample data set to obtain a target prediction model.

[0014] Optionally, the sample data set includes a labeled training set, a test set, and a validation set, and the step of generating the sample data set according to the feedback information includes:

[0015] Associating the feedback information with historical pickup information of each user in the first user group to obtain a first sample set;

[0016] The first sample set is divided according to the time period corresponding to the feedback information to obtain a labeled training set, a test set, and a validation set.

[0017] Optionally, the sample data set further includes an unlabeled training set, and the step of iteratively training a preset basic prediction model using the sample data set to obtain a target prediction model includes:

[0018] Obtaining historical pickup information for a second user group, and associating the historical pickup information for the second user group with the account information of each user in the second user group to generate an unlabeled training set, wherein the second user group consists of users who did not generate feedback information after picking up their parcels;

[0019] Performing unsupervised learning on the unlabeled training set to obtain a first data set, and performing non-repetitive random sampling and combination on the first data set by columns to obtain a second data set;

[0020] Merging the second data set with the labeled training set to obtain a target training set;

[0021] The target training set, the test set and the validation set are used to perform supervised learning on the preset basic prediction model to obtain the target prediction model.

[0022] Optionally, the step of performing supervised learning on a preset basic prediction model using the target training set, the test set, and the validation set to obtain a target prediction model includes:

[0023] Performing supervised learning on a preset basic prediction model using the target training set to obtain a first prediction model, wherein the first prediction model includes a plurality of;

[0024] The test set is used to test each of the first prediction models, and the prediction results of the first prediction models are evaluated using the Gini coefficient, and a target prediction model is determined from each of the first prediction models based on the Gini coefficient.

[0025] Optionally, the step of outputting a warning message when it is detected that the number of packages to be picked up corresponding to the package information to be picked up exceeds the load capacity of the user includes:

[0026] When it is detected that the number of packages to be picked up corresponding to the package to be picked up information exceeds the load capacity of the user, the total contribution value of each feature information in the account information and the weather information is calculated respectively;

[0027] Determining the target feature with the largest total contribution according to the total contribution value;

[0028] Output warning prompt information according to the target characteristics.

[0029] In addition, to achieve the above-mentioned purpose, the present invention further provides a pickup volume warning device, which includes:

[0030] A data acquisition module is used to obtain and save weather information, and when the user's account information is detected, determine the information of the package to be picked up based on the account information;

[0031] a pickup quantity prediction module, configured to predict, based on the account information and the weather information, whether the pickup quantity indicated by the to-be-pickup package information exceeds the user's load capacity;

[0032] The early warning prompt module is used to output early warning prompt information when the prediction result shows that the pickup quantity indicated by the to-be-picked package information exceeds the load capacity of the user.

[0033] In addition, to achieve the above-mentioned purpose, the present invention also provides a terminal device, which includes: a memory, a processor, and a pickup volume warning program stored on the memory and runnable on the processor. When the pickup volume warning program is executed by the processor, the steps of the pickup volume warning method as described above are implemented.

[0034] In addition, to achieve the above-mentioned purpose, the present invention also provides a computer-readable storage medium, on which a pickup volume warning program is stored. When the pickup volume warning program is executed by the processor, the steps of the pickup volume warning method as described above are implemented.

[0035] In addition, to achieve the above-mentioned purpose, the present invention also provides a computer program product, including a computer program, which, when executed by a processor, implements the steps of the above-mentioned pickup volume warning method.

[0036] Embodiments of the present invention provide a pickup volume warning method, apparatus, terminal device, and storage medium. In the prior art, due to the loose integration of prediction models with service characteristics, overfitting is prone to occur, resulting in inaccurate predictions of a user's single pickup volume, which can easily affect the user's normal walking after picking up the packages. Compared to the prior art, the embodiments of the present invention obtain and store weather information. When a user's account information is detected, the system determines pending package information based on the account information. Based on the account information and weather information, it predicts whether the pickup volume indicated by the pending package information exceeds the user's load capacity. If the prediction result shows that the pickup volume indicated by the pending package information exceeds the user's load capacity, a warning message is output. This can combine weather and user factors to predict a user's single pickup volume, improving the accuracy of the prediction. When the pending packages exceed the user's load capacity, a warning message is output to prompt the user to reduce the number of packages to be picked up, preventing the user from picking up too many packages at once, which would affect their normal walking, thereby improving the user experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 A schematic diagram of the hardware structure of a terminal device according to an embodiment of the present invention;

[0038] Figure 2 This is a flow chart of the first embodiment of the pickup volume warning method of the present invention;

[0039] Figure 3 This is a flow chart of model training for the second embodiment of the pickup volume early warning method of the present invention;

[0040] Figure 4 This is a schematic diagram of the functional modules of an embodiment of the pickup volume warning device of the present invention.

[0041] The purpose, features and advantages of the present invention will be further described with reference to the accompanying drawings and in conjunction with the embodiments. DETAILED DESCRIPTION

[0042] It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0043] In the subsequent description, suffixes such as "module," "component," or "unit" used to represent elements are only used to facilitate the description of the present invention and have no specific meaning. Therefore, "module," "component," or "unit" can be used interchangeably.

[0044] The pickup volume warning terminal (also called terminal, device or terminal device) of the embodiment of the present invention can be a PC, or a mobile terminal device with display and data processing functions such as a smart phone, tablet computer and portable computer.

[0045] like Figure 1 As shown, the terminal may include: a processor 1001, such as a CPU, a network interface 1004, a user interface 1003, a memory 1005, and a communication bus 1002. Among them, the communication bus 1002 is used to realize the connection and communication between these components. The user interface 1003 may include a display screen (Display), an input unit such as a keyboard (Keyboard), and the user interface 1003 may also include a standard wired interface and a wireless interface. The network interface 1004 may optionally include a standard wired interface and a wireless interface (such as a WI-FI interface). The memory 1005 may be a high-speed RAM memory or a stable memory (non-volatile memory), such as a disk memory. The memory 1005 may also be a storage device independent of the aforementioned processor 1001.

[0046] Optionally, the terminal may also include a camera, an RF (Radio Frequency) circuit, a sensor, an audio circuit, a WiFi module, and the like. Among them, sensors include light sensors, motion sensors, and other sensors. Specifically, the light sensor may include an ambient light sensor and a proximity sensor, wherein the ambient light sensor may adjust the brightness of the display screen according to the brightness of the ambient light, and the proximity sensor may turn off the display screen and / or backlight when the mobile terminal is moved to the ear. As a type of motion sensor, the gravity acceleration sensor can detect the magnitude of acceleration in all directions (generally three axes), and can detect the magnitude and direction of gravity when stationary. It can be used for applications that identify the posture of the mobile terminal (such as horizontal and vertical screen switching, related games, magnetometer posture calibration), vibration recognition related functions (such as pedometer, tapping), etc.; of course, the mobile terminal can also be configured with other sensors such as gyroscopes, barometers, hygrometers, thermometers, infrared sensors, etc., which will not be repeated here.

[0047] Those skilled in the art will understand that Figure 1 The terminal structure shown in the figure does not constitute a limitation to the terminal, and may include more or fewer components than shown in the figure, or combine certain components, or arrange the components differently.

[0048] like Figure 1 As shown, the memory 1005 as a computer-readable storage medium may include an operating system, a network communication module, a user interface module, and a pickup volume warning program.

[0049] exist Figure 1 In the terminal shown, the network interface 1004 is mainly used to connect to the backend server and communicate data with the backend server; the user interface 1003 is mainly used to connect to the client (user end) and communicate data with the client; and the processor 1001 can be used to call the pickup volume warning program stored in the memory 1005. When the pickup volume warning program is executed by the processor, the operations in the pickup volume warning method provided in the following embodiment are implemented.

[0050] Based on the above-mentioned device hardware structure, an embodiment of the pickup volume warning method of the present invention is proposed.

[0051] Reference Figure 2 In a first embodiment of the pickup volume early warning method of the present invention, the pickup volume early warning method includes:

[0052] Step S10, obtaining and saving weather information, and when the user's account information is detected, determining the information of the package to be picked up according to the account information;

[0053] In this embodiment, the pickup volume warning method can be implemented in mobile terminals such as PCs and tablets and applied to smart express lockers (hereinafter referred to as express lockers). It can combine multiple factors such as user information and weather information to predict the user's pickup volume, so that when the user's to-be-picked packages exceed its load capacity, a warning prompt message is output to prompt the user to reduce the single pickup volume.

[0054] Specifically, the express locker has communication and storage functions, which can save the user's pickup information and weather information, etc., and can also interact with the server to realize the pickup volume prediction function. First, the weather information is obtained and saved. The weather information can be obtained by obtaining weather forecast information from the server, or by directly obtaining the weather information of the express locker location through the camera set on the smart express locker. When obtaining weather information, it can be obtained once every preset time period, and the preset time period can be one minute or one hour, which is not specifically limited here. The weather information can be saved locally in the express locker or saved to the server, which is also not specifically limited here.

[0055] When the user's account information is detected, the information of the parcels to be picked up is determined based on the account information. The information to be picked up includes the number of parcels to be picked up, the volume and weight of each parcel to be picked up, etc. The detection of the user's account information can be when the courier deposits the parcel into the express locker during the delivery process. It is known that when the courier deposits the parcel into the express locker, the courier needs to enter the recipient information (i.e., user information). The user information includes at least the recipient's name and contact information. Based on the recipient's name and / or contact information, it can be associated with the user's account information stored locally in the express locker. The detection of the user's account information can also be when the user logs into the express locker to pick up the parcel. It is known that when sending or receiving parcels, the user generally scans the QR code on the express locker display screen through a terminal such as a mobile phone, or enters his or her account information through the virtual button on the express locker display screen to log in to the express locker's user account. It should be noted that in this embodiment, each user has a unique identification code, and each identification code corresponds to an account information. The same express locker can have account information for multiple users, and the account information of each user is unique and non-repetitive. Different users can create accounts for different express lockers, or the same user account can log in to different lockers, without specific restrictions here. When the same user's account information can log in to different lockers, the user's pending package information in different lockers can be used to predict whether the user's pickup volume exceeds the locker's load capacity, thereby improving prediction accuracy.

[0056] Step S20, predicting the user's load capacity based on the account information and the weather information;

[0057] After obtaining weather information and user account information, the user's load capacity is predicted based on the user's account information and weather information. It is clear that the user's load capacity is affected by many factors, such as gender, age, weather conditions at the time of pickup, and transportation used for pickup. Based on the user's account information, the user's historical pickup data stored locally or on the server can be linked to the current weather information to predict the user's load capacity.

[0058] Step S30: When it is detected that the number of packages to be picked up corresponding to the package information to be picked up exceeds the load capacity of the user, a warning prompt message is output.

[0059] When it is detected that the number of pending packages corresponding to the pending package information exceeds the user's load capacity, a warning message is output, prompting the user to reduce the number of packages to be picked up to avoid affecting normal activities such as walking. Exceeding the user's load capacity may mean that the number of packages exceeds the maximum number the user can pick up, and / or the weight of the packages exceeds the weight the user can carry, and / or the volume of the packages exceeds the volume the user can carry. As is known, express delivery lockers are generally equipped with lockers of different capacities, such as large, medium, and small. After obtaining the number of pending packages from the user, the volume of each package can be preliminarily estimated based on the number of large, medium, and small compartment lockers occupied by the pending packages. The weight of the packages can be manually entered by the courier when dropping off the packages, or a weighing device can be installed in the express delivery locker to obtain the weight of each package, thereby determining the total weight of the user's pending packages. The number, volume, and weight of the user's pending packages are compared with the user's load capacity. If the number exceeds the user's load capacity, a warning message is output, prompting the user to reduce the number of packages to be picked up at a time or use transportation. This warning message can be displayed on the display screen of the express cabinet when the user picks up the package, such as "To avoid picking up too many packages and affecting walking, it is recommended to reduce the number of packages picked up", or it can be pushed to the user's mobile phone or other terminals, such as "There are a large number of packages to be picked up (or large in size or heavy in weight). To ensure your normal movement, it is recommended to prepare transportation in advance or reduce the number of packages picked up at a single time", etc., without specific limitations here.

[0060] Furthermore, a warning message can be output when the predicted probability of the number of pending packages exceeding the user's load capacity exceeds a set threshold. For example, if the threshold is set to 0.98, a warning message is output when the probability of the number of pending packages exceeding the user's load capacity exceeds 0.98. After the warning message is output, if the user confirms a request to reduce the number of packages to be picked up, the system recommends packages that can be picked up to the user based on the predicted user's load capacity. Furthermore, if the user has a large number of pending packages, the system optimizes the user's pickup plan, ensuring that the number of package pickups is minimized without affecting the user's normal activities, thereby improving the user experience.

[0061] In this embodiment, by acquiring and storing weather information, when user account information is detected, pending package information is determined based on the account information. Based on the account information and weather information, a prediction is made as to whether the pickup quantity indicated by the pending package information exceeds the user's load capacity. If the prediction indicates that the pickup quantity indicated by the pending package information exceeds the user's load capacity, a warning message is output. This can combine weather and user factors to predict a user's single pickup quantity, improving the accuracy of the prediction. Consequently, when the number of pending packages exceeds the user's load capacity, a warning message is output to prompt the user to reduce the number of packages to be picked up, preventing the user from picking up too many packages at once, which would affect their normal travel, thereby improving the user experience.

[0062] Furthermore, based on the above-mentioned embodiment of the present invention, a second embodiment of the pickup volume warning method of the present invention is proposed.

[0063] This embodiment is a refinement of step S20 in the first embodiment, and includes the following steps:

[0064] Step S201: inputting the account information and the weather information into a target prediction model, and extracting target feature information from the account information and the weather information using the target prediction model;

[0065] Step S202, predicting the user's load capacity based on the target feature information, wherein the target prediction model is obtained by iteratively training a preset basic prediction model using a sample data set, and the sample data set is generated based on each user's account information, weather information at the time of pickup, and feedback information after pickup.

[0066] In this embodiment, when predicting a user's load capacity, the user's account information and acquired weather information are input into a trained target prediction model for prediction. Specifically, the user's account information and acquired weather information are input into the target prediction model, and then the user's account information is linked to the user's saved historical pickup information and the weather information at the time of pickup. The target prediction model is used to extract target feature information from the account information and weather information, and the extracted target feature information is matched with the historical pickup information to predict the user's load capacity. The target prediction model is obtained by iteratively training a preset basic prediction model using a sample data set. The sample data set is generated based on each user's account information, each user's historical weather information when picking up the package, and feedback information after the pickup. The user's feedback information is the information provided by the user after the pickup whether it affects walking. Based on the weather information from the user's past pickup times, the user's historical pickup volume, and feedback on whether it affects walking, the collected information is correlated through model training to identify correlations. This allows the trained prediction model to predict the user's load capacity based on their account information and weather information when the user picks up their next package. The prediction target is how many boxes of items the user will be overloaded with, thereby determining whether the user can carry all the packages currently waiting to be picked up. Furthermore, the user's feedback information can be obtained through an application configured on a terminal such as a mobile phone, or it can be feedback information from the user's previous pickup at a courier, without specific limitations here.

[0067] Furthermore, before inputting the user's account information and the obtained weather information into the behind-the-scenes prediction model, the prediction model needs to be trained. First, a sample dataset for training is constructed:

[0068] Step A1: obtaining feedback information from a first user group after picking up a package, where the first user group is composed of users who generate feedback information after picking up a package;

[0069] Step A2: generating a sample data set according to the feedback information, and iteratively training a preset basic prediction model using the sample data set to obtain a target prediction model.

[0070] Feedback from the first user group after picking up their parcels is obtained. This feedback information is provided by each user in the first user group regarding whether their trip was impacted. The first user group consists of users who provided feedback after picking up their parcels. A sample dataset is generated based on the feedback from each user in the first user group. This sample dataset is used to iteratively train a pre-set basic prediction model to obtain a target prediction model.

[0071] Furthermore, the generated sample dataset includes a labeled training set, a validation set, and a test set. The specific construction process of the sample dataset includes:

[0072] Step B1: Associating the feedback information with the account information of each user in the first user group and the weather information at the time of pickup to obtain a first sample set;

[0073] Step B2: Divide the first sample set according to the time period corresponding to the feedback information to obtain a labeled training set, a test set, and a validation set.

[0074] After receiving user feedback, the user's feedback is associated with the previously stored historical pickup information of the user to obtain a first sample set. This first sample set is then divided according to the time period corresponding to the received feedback information to obtain labeled training, test, and validation sets. The user's historical pickup information includes account information, weather information at the time of pickup, and the number of pickups.

[0075] Furthermore, the sample data set is a wide table of data composed of feature fields related to user account information, weather information, etc. The features in the wide table include at least: user unique identification code, pickup time (date and time of pickup, which can be accurate to seconds or minutes), weather type of the express cabinet location (sunny or cloudy, etc.), temperature of the express cabinet location, precipitation at the express cabinet location, user gender, means of transportation used for pickup, whether there is an elevator at the destination, precipitation within the first time period after pickup (such as 10 minutes), precipitation within the second time period after pickup (such as 30 minutes), area type of the express cabinet location (such as commercial area or residential area, etc.), number of large box grids occupied, number of medium box grids occupied, number of small box grids occupied, number of pickups within a preset time period (such as 90 days), whether it is within the load range (feedback information on whether the user can bear the volume or weight of the package), etc.

[0076] The collected data wide table is divided according to the time period corresponding to the user feedback information. For example, the user feedback information collected in the Nth time period is associated with the user's historical pickup information stored in advance to form a wide table. The model prediction target is marked according to the user's feedback information to obtain the test set a. The user feedback information collected in the N+1th time period is associated with the user's historical pickup information stored in advance to form a wide table. The model prediction target is marked according to the user's feedback information to obtain the verification set b. The user feedback information collected in the N+2th time period is associated with the user's historical pickup information stored in advance to form a wide table. The model prediction target is marked according to the user's feedback information to obtain the labeled training set c.

[0077] Furthermore, the sample data set used for training the prediction model also includes an unlabeled training set. The steps of iteratively training the preset basic prediction model to obtain the target prediction model include:

[0078] Step C1: Obtain historical pickup information of a second user group, and associate the historical pickup information of the second user group with the account information of each user in the second user group to generate an unlabeled training set, wherein the second user group consists of users who did not generate feedback information after picking up their parcels;

[0079] Step C2, performing unsupervised learning on the unlabeled training set to obtain a first data set, and performing non-repetitive random sampling and combination on the first data set by column to obtain a second data set;

[0080] Step C3, merging the second data set with the labeled training set to obtain a target training set;

[0081] Step C4: Using the target training set, the test set, and the validation set, supervised learning is performed on the preset basic prediction model to obtain a target prediction model.

[0082] Historical pickup information for a second user group is obtained and associated with the account information of each user in the second user group to generate an unlabeled training set. The second user group consists of users who did not provide feedback after picking up their parcels. Specifically, the unlabeled training set can be generated by filtering the relational database storing user pickup information to identify users who never provided feedback. This unlabeled training set is essentially the same as the aforementioned wide data table, except that the field value for "whether within the load range" is left blank, indicating that the user did not provide feedback after picking up their parcel.

[0083] Unsupervised learning is performed on the unlabeled training set. An unsupervised machine learning algorithm is used to label the prediction targets into two categories, resulting in a corresponding dataset d. Unsupervised learning is then combined with periodic data features to find the optimal combination of labeled data. Expert experience is then used to label the two categories in dataset d as "walking difficulty" and "walking without difficulty," respectively, to obtain the first dataset, dataset e. An unsupervised machine learning algorithm is used to predict "whether the person can carry a certain volume or weight." Through clustering, these categories are labeled as 0 or 1, resulting in dataset d. Expert experience modifies these 0s and 1s to "yes" and "no," resulting in dataset e. Based on the number of rows and row numbers in dataset e, columns in dataset e are randomly sampled without duplication. The sampled data are combined to obtain a second dataset, such as f1, f2, f3, ..., fn. These second datasets f1, f2, f3, ..., fn are combined with the labeled training set c to obtain the target training set, such as f1c, f2c, f3c, ..., fnc.

[0084] The target training set, test set, and validation set are used to perform supervised learning on the preset basic prediction model to obtain the target prediction model. Specifically, f1c, f2c, f3c, ..., fnc are used as training sets to perform supervised learning on the preset basic prediction model to obtain the target prediction model.

[0085] Furthermore, the steps of using the target training set, test set, and validation set to conduct supervised learning on the preset basic prediction model to obtain the target prediction model include:

[0086] Step D1, using the target training set to perform supervised learning on a preset basic prediction model to obtain a first prediction model, wherein the first prediction model includes multiple;

[0087] Step D2: Test each of the first prediction models using the test set, evaluate the prediction results of the first prediction models using the Gini coefficient, and determine a target prediction model from each of the first prediction models based on the Gini coefficient.

[0088] Reference Figure 3 , Figure 3 This is a flow chart of iterative training of the preset basic prediction model in this embodiment. Figure 3 In [1], collected user feedback is associated with previously stored historical user pickup information to form a wide table. This table is then divided into datasets a, b, and c based on the time period during which the feedback was collected. A test set is a table, b is a validation set, and c is a labeled training set. Historical pickup information for users who provided feedback is then extracted from a relational database to form a wide table without feedback labels, serving as the unlabeled training set. Using an unsupervised machine learning algorithm, the prediction targets in the unlabeled training set are labeled into two categories, representing whether the user can afford to pick up the package, resulting in dataset d. In dataset d, the two categories are labeled "0" and "1," respectively. Using expert experience, the two "0" and "1" categories in dataset d are labeled as "yes" and "no," resulting in dataset e. In dataset e, "yes" indicates that the user can afford to pick up the package and has no difficulty walking, while "no" indicates that the user cannot afford to pick up the package and has difficulty walking.

[0089] Furthermore, the entire column data of the data set e is randomly extracted and combined without repetition according to the number of rows and row numbers to obtain unlabeled training sets f1, f2, f3...fn. The data sets f1, f2, f3...fn are merged with the data set c in turn to obtain target training sets f1c, f2c, f3c...fnc. Then, f1c, f2c, f3c...fnc are used as training sets, the data set a is used as the test set, and the data set b is used as the validation set. After a supervised machine learning algorithm, multiple first prediction models M1, M2, M3...Mn are obtained. Then, the models M1, M2, M3...Mn are used to predict the test set a respectively, and the prediction results are evaluated using the Gini coefficient. The model with the largest Gini coefficient is taken as the target prediction model.

[0090] In this embodiment, a combination of unsupervised learning and supervised learning is adopted. The training set, test set, and validation set are divided by time series. Combined with the data characteristics of periodic regularities, the best combination of data amounts labeled by the unsupervised algorithm is found as the training set. The data set screened in this way is used as the training set to train the model actually used, which can effectively prevent the model from overfitting. In the process of model training, combining expert experience can also make up for the limitations of simply using unsupervised machine learning algorithm clustering, thereby improving the prediction accuracy of the model.

[0091] Furthermore, based on the first and / or second embodiments described above, a third embodiment of the pickup volume warning method of the present invention is proposed. This embodiment is a refinement of step S30 in the first embodiment described above, and mainly includes:

[0092] Step S301, when it is detected that the number of packages to be picked up corresponding to the package to be picked up information exceeds the load capacity of the user, calculating the total contribution value of each feature information in the account information and the weather information respectively;

[0093] Based on the above embodiment, this embodiment uses a target prediction model to pre-store a user's load capacity. When it is detected that the number of packages to be picked up exceeds the user's load capacity, a warning message is output to prompt the user to reduce the number of packages to be picked up or use transportation when picking up the packages. In this embodiment, different types of warning messages are output based on the main factors that affect the user's load capacity.

[0094] Specifically, when it is detected that the number of packages to be picked up exceeds the user's load capacity, the aggregate contribution value of each feature information in the user's account information and the acquired weather information to the model prediction is calculated. The feature information in the user's account information includes at least the user code (i.e., the user's unique identification code) and the user's historical pickup times. The feature information in the weather information includes at least the weather type, temperature, and precipitation. The precipitation includes the precipitation amount within different time periods after the user picks up the package. Specifically, the feature information in the account information and weather information can be the same as the fields in the wide table in the sample dataset.

[0095] Specifically, when calculating the total contribution value of each feature information, the contribution values ​​of the feature information extracted according to different dimensions are summed. Specifically, from the user dimension, the contribution values ​​of each feature information related to the user are summed to obtain the total contribution value of the user dimension. From the weather dimension, the contribution values ​​of each feature information related to the weather information are summed to obtain the total contribution value of the weather dimension. It can be seen that when calculating the total contribution value, the various feature information in the above wide table can also be divided from other more dimensions, which will not be repeated here.

[0096] Step S302, determining the target feature with the largest total contribution according to the total contribution value;

[0097] Based on the calculated total contribution values ​​for different dimensions, the target feature with the highest contribution is determined. Essentially, this means identifying the primary factor influencing user load capacity. For example, the primary factor affecting user load capacity could be the user's physical fitness or the weather. Taking weather as an example, the contribution values ​​of all weather-related fields in the aforementioned wide table entry data are summed to obtain a total value, Y. If Y is greater than the total contribution value, Z, of all other fields, then weather is the primary factor influencing user load capacity.

[0098] Step S303: outputting warning information according to the target characteristics.

[0099] Warning prompt information is output based on the target feature with the largest total contribution value. In this embodiment, different types of prompt information are output from different dimensions depending on the main factors affecting the user's load capacity. For example, when the total contribution value of weather-related features is the largest, the main factor affecting the user's load capacity is the weather, and the warning prompt information output and displayed may be "Considering weather factors, it is recommended to reduce the number of pickups!" When the main factor affecting the user's load capacity is the user's own physical fitness, such as if there are large or heavy packages to be picked up, the warning prompt information output and displayed may be "Considering your physical fitness and other factors, it is recommended to reduce the number of pickups!" Specifically, the warning prompt information output and displayed can be adaptively adjusted based on the different feature information with the largest total contribution value, which will not be repeated here.

[0100] In this embodiment, when it is detected that the number of packages to be picked up exceeds the user's load capacity, the main factors affecting the user's load capacity are determined by calculating the total contribution value of different feature information, and different types of early warning prompt information are output according to different factors, thereby improving the flexibility of the early warning prompt. At the same time, after predicting the load capacity of the user after picking up the package, the early warning prompt information is issued in a timely manner, which can prevent the user from picking up too many packages at one time and affecting his walking, thereby enhancing the service capabilities of the smart express cabinet and improving the user experience.

[0101] In addition, refer to Figure 4 The embodiment of the present invention further provides a pickup volume warning device, the pickup volume warning device comprising:

[0102] The data acquisition module 10 is used to obtain and store weather information and, when the user's account information is detected, determine the information of the package to be picked up based on the account information;

[0103] a pickup quantity prediction module 20, configured to predict, based on the account information and the weather information, whether the pickup quantity indicated by the to-be-pickup package information exceeds the user's load capacity;

[0104] The early warning prompt module 30 is configured to output an early warning prompt message when the prediction result shows that the pickup quantity indicated by the to-be-collected parcel information exceeds the load capacity of the user.

[0105] Optionally, the pickup quantity prediction module 20 is further configured to:

[0106] inputting the account information and the weather information into a target prediction model, and extracting target feature information from the account information and the weather information using the target prediction model;

[0107] The user's load capacity is predicted based on the target feature information, wherein the target prediction model is obtained by iteratively training a preset basic prediction model using a sample data set, and the sample data set is generated based on each user's account information, weather information at the time of pickup, and feedback information after pickup.

[0108] Optionally, the pickup volume prediction device further includes a model training module, which is used to:

[0109] Obtaining feedback information from a first user group after picking up a package, where the first user group is composed of users who generate feedback information after picking up a package;

[0110] A sample data set is generated according to the feedback information, and a preset basic prediction model is iteratively trained using the sample data set to obtain a target prediction model.

[0111] Optionally, the sample data set includes a labeled training set, a test set, and a validation set, and the model training module is further used to:

[0112] Associating the feedback information with historical pickup information of each user in the first user group to obtain a first sample set;

[0113] The first sample set is divided according to the time period corresponding to the feedback information to obtain a labeled training set, a test set, and a validation set.

[0114] Optionally, the sample data set further includes an unlabeled training set, and the model training module is further configured to:

[0115] Obtaining historical pickup information for a second user group, and associating the historical pickup information for the second user group with the account information of each user in the second user group to generate an unlabeled training set, wherein the second user group consists of users who did not generate feedback information after picking up their parcels;

[0116] Performing unsupervised learning on the unlabeled training set to obtain a first data set, and performing non-repetitive random sampling and combination on the first data set by columns to obtain a second data set;

[0117] Merging the second data set with the labeled training set to obtain a target training set;

[0118] The target training set, the test set and the validation set are used to perform supervised learning on the preset basic prediction model to obtain the target prediction model.

[0119] Optionally, the model training module is further used to:

[0120] Performing supervised learning on a preset basic prediction model using the target training set to obtain a first prediction model, wherein the first prediction model includes a plurality of;

[0121] The test set is used to test each of the first prediction models, and the prediction results of the first prediction models are evaluated using the Gini coefficient, and a target prediction model is determined from each of the first prediction models based on the Gini coefficient.

[0122] Optionally, the early warning module 30 is further configured to:

[0123] When it is detected that the number of packages to be picked up corresponding to the package to be picked up information exceeds the load capacity of the user, the total contribution value of each feature information in the account information and the weather information is calculated respectively;

[0124] Determining the target feature with the largest total contribution according to the total contribution value;

[0125] Output warning prompt information according to the target characteristics.

[0126] In addition, an embodiment of the present invention also proposes a computer-readable storage medium, on which a pickup volume warning program is stored. When the pickup volume warning program is executed by a processor, the operations in the pickup volume warning method provided in the above embodiment are implemented.

[0127] In addition, an embodiment of the present invention further provides a computer program product, including a computer program, which, when executed by a processor, implements the operations in the pickup volume warning method provided in the above embodiment.

[0128] The various embodiments of the device, computer program product, and computer-readable storage medium of the present invention may refer to the various embodiments of the pickup volume warning method of the present invention, and will not be repeated here.

[0129] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity / operation / object from another entity / operation / object, and do not necessarily require or imply any actual relationship or order between these entities / operations / objects; the terms "include", "comprise", or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article, or system that includes a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article, or system. In the absence of further limitations, an element defined by the phrase "includes a ..." does not exclude the presence of other identical elements in the process, method, article, or system that includes the element.

[0130] As for the device embodiment, since it is basically similar to the method embodiment, the description is relatively simple. For relevant parts, please refer to the partial description of the method embodiment. The device embodiment described above is merely illustrative, and the units described as separate components may or may not be physically separated. Some or all of the modules can be selected according to actual needs to achieve the purpose of the present invention. Those skilled in the art can understand and implement it without paying any creative work.

[0131] The serial numbers of the above embodiments of the present invention are for description only and do not represent the advantages or disadvantages of the embodiments.

[0132] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus the necessary general hardware platform, and of course can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention is essentially or the part that contributes to the existing technology can be embodied in the form of a software product. The computer software product is stored in a storage medium (such as ROM / RAM, disk, CD) as described above, and includes a number of instructions for enabling a terminal device (which can be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the pickup volume warning method described in each embodiment of the present invention.

[0133] The above are only preferred embodiments of the present invention and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention description and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A pickup volume warning method, applied to smart express lockers, characterized by: The pickup volume early warning method comprises the following steps: Acquire and save weather information, and when the user's account information is detected, determine the information of the package to be picked up based on the account information; inputting the account information and the weather information into a target prediction model, and extracting target feature information from the account information and the weather information using the target prediction model; Predicting the user's load capacity based on the target feature information, and outputting a warning prompt message when it is detected that the number of packages to be picked up corresponding to the package information to be picked up exceeds the user's load capacity; Before inputting the account information and the weather information into the target prediction model, the method further includes: Obtaining feedback information from a first user group after picking up a package, where the first user group is composed of users who generate feedback information after picking up a package; Associating the feedback information with historical pickup information of each user in the first user group to obtain a first sample set; Dividing the first sample set according to the time period corresponding to the feedback information to obtain a sample data set including a labeled training set, a test set, and a validation set; Obtaining historical pickup information for a second user group, and associating the historical pickup information for the second user group with the account information of each user in the second user group to generate an unlabeled training set, wherein the second user group consists of users who did not generate feedback information after picking up their parcels; Performing unsupervised learning on the unlabeled training set to obtain a first data set, and performing non-repetitive random sampling and combination on the first data set by columns to obtain a second data set; Merging the second data set with the labeled training set to obtain a target training set; The target training set, the test set and the validation set are used to perform supervised learning on the preset basic prediction model to obtain the target prediction model.

2. The pickup volume warning method according to claim 1, characterized in that: The step of performing supervised learning on a preset basic prediction model using the target training set, the test set, and the validation set to obtain a target prediction model includes: Performing supervised learning on a preset basic prediction model using the target training set to obtain a first prediction model, wherein the first prediction model includes a plurality of; The test set is used to test each of the first prediction models, and the prediction results of the first prediction models are evaluated using the Gini coefficient, and a target prediction model is determined from each of the first prediction models based on the Gini coefficient.

3. The pickup volume warning method according to claim 1, characterized in that: The step of outputting a warning prompt message when it is detected that the to-be-picked packages corresponding to the to-be-picked package information exceed the load capacity of the user includes: When it is detected that the number of packages to be picked up corresponding to the package to be picked up information exceeds the load capacity of the user, calculating the total contribution value of each feature information in the account information and the weather information respectively; Determining the target feature with the largest total contribution according to the total contribution value; Output warning prompt information according to the target characteristics.

4. A pickup volume warning device, characterized in that: The pickup volume warning device includes: A data acquisition module is used to obtain and save weather information, and when the user's account information is detected, determine the information of the package to be picked up based on the account information; The data acquisition module is also used to obtain feedback information from a first user group after picking up a package, the first user group being composed of users who generate feedback information after picking up a package; associating the feedback information with historical package pickup information of each user in the first user group to obtain a first sample set; dividing the first sample set according to the time period corresponding to the feedback information to obtain a sample data set including a labeled training set, a test set, and a validation set; obtaining historical package pickup information of a second user group, and associating the historical package pickup information of the second user group with the account information of each user in the second user group to generate an unlabeled training set, wherein the second user group is composed of users who do not generate feedback information after picking up a package; performing unsupervised learning on the unlabeled training set to obtain a first data set, and performing non-repetitive random sampling and combination on the first data set by column to obtain a second data set; merging the second data set with the labeled training set to obtain a target training set; using the target training set, the test set, and the validation set to perform supervised learning on a preset basic prediction model to obtain a target prediction model; A pickup quantity prediction module is configured to input the account information and the weather information into a target prediction model, extract target feature information from the account information and the weather information using the target prediction model, and predict the user's load capacity based on the target feature information; and an early warning prompt module is configured to output an early warning prompt message when the prediction result shows that the pickup quantity indicated by the to-be-picked package information exceeds the user's load capacity.

5. A terminal device, characterized in that: The terminal device includes: a memory, a processor, and a pickup volume warning program stored in the memory and executable on the processor. When the pickup volume warning program is executed by the processor, the steps of the pickup volume warning method as described in any one of claims 1 to 3 are implemented.

6. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a pickup volume warning program, which, when executed by a processor, implements the steps of the pickup volume warning method according to any one of claims 1 to 3.

Images