Data processing and analysis method and device

Abstract

The present invention discloses a method and device for data processing and analysis, and relates to the technical field of computer technology. A specific implementation of the method includes: constructing a training database and a prediction database related to users; extracting behavioral feature identification data related to the users from the training database; fitting the behavioral feature identification data to obtain an optimization objective function related to the user's shopping time; establishing a first prediction model based on the optimization objective function; and using the first prediction model to obtain the predicted shopping time related to the user based on the prediction database. This implementation has the technical effect of improving the pertinence and accuracy of the product recommendation scheme, and further provides effective parameters for other optimization methods based on e-commerce.

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and in particular to a method and device for data processing and analysis. Background Art

[0002] In e-commerce scenarios, how to use data-driven to provide more helpful information to our decision makers has been a hot topic in recent years. With the accumulation of big data, we have been able to achieve data-driven in many scenarios through big data plus machine learning technology. For the e-commerce field, data-driven can enable us to better understand our users, so as to develop more sophisticated marketing plans for our users, such as product recommendations for specific users.

[0003] In the process of implementing the present invention, the inventors found that there are at least the following problems in the prior art:

[0004] Because the existing technology lacks the ability to estimate the user's next shopping time when formulating marketing plans for online shoppers, it is impossible to make an overall plan and control the user's next shopping process within a limited time, so the targeting and accuracy of the marketing plan needs to be strengthened. Summary of the invention

[0005] In view of this, an embodiment of the present invention provides a method and device for data processing and analysis, which can predict the duration of a user's next shopping trip, thereby making more targeted customized product recommendations and other services within the user's limited shopping time.

[0006] To achieve the above objective, according to one aspect of an embodiment of the present invention, a method for data processing and analysis is provided.

[0007] The data processing and analysis method according to an embodiment of the present invention includes:

[0008] Construct user-related training database and prediction database;

[0009] Extracting behavioral feature identification data related to the user from the training database;

[0010] Fitting the behavior feature recognition data to obtain an optimization objective function related to the user's shopping time;

[0011] Establishing a first prediction model based on the optimization objective function; and

[0012] The first prediction model is used to obtain a predicted shopping duration associated with the user based on the prediction database.

[0013] Optionally, the behavior feature identification data includes one or more of the following:

[0014] User basic features, user cluster group features, user product interaction features, and user text features.

[0015] Optionally, in the user-product interaction feature, a time-related ID code is performed on the product related to the user.

[0016] Optionally, the user cluster group feature is a single-layer cluster feature or a multi-layer cluster feature related to the user.

[0017] Optionally, obtaining a predicted shopping duration related to the user based on the prediction database using the first prediction model further includes:

[0018] Construct a user-related authentication database;

[0019] Using the verification database to verify and adjust the first prediction model to obtain a second prediction model; and

[0020] The second prediction model is used to obtain a predicted shopping duration related to the user based on the prediction database.

[0021] According to one aspect of an embodiment of the present invention, a device for data processing and analysis is provided.

[0022] The data processing and analysis device according to an embodiment of the present invention includes:

[0023] A database construction module, used to construct a training database and a prediction database related to users;

[0024] A feature group extraction module, used to extract behavioral feature identification data related to the user from the training database;

[0025] An optimization objective function fitting module is used to fit the behavior feature recognition data to obtain an optimization objective function related to the user's shopping time;

[0026] A model building module, used to build a first prediction model based on the optimization objective function; and

[0027] A shopping duration prediction module is used to obtain a predicted shopping duration related to the user based on the prediction database using the first prediction model.

[0028] Optionally, the behavior feature identification data includes one or more of the following:

[0029] User basic features, user cluster group features, user product interaction features, and user text features.

[0030] Optionally, the data processing and analysis device further comprises:

[0031] The ID encoding module is used to perform time-related ID encoding on the commodities related to the user in the user-commodity interaction feature.

[0032] Optionally, the user cluster group feature is a single-layer cluster feature or a multi-layer cluster feature related to the user.

[0033] Optionally,

[0034] The database construction module is also used to construct a verification database related to the user;

[0035] The model building module is further configured to use the verification database to verify and adjust the first prediction model to obtain a second prediction model; and

[0036] The shopping duration prediction module is further used to obtain the predicted shopping duration related to the user based on the prediction database using the second prediction model.

[0037] According to another aspect of an embodiment of the present invention, an electronic device for data processing and analysis is provided.

[0038] An electronic device for data processing and analysis according to an embodiment of the present invention includes:

[0039] one or more processors;

[0040] a storage device for storing one or more programs,

[0041] When the one or more programs are executed by the one or more processors, the one or more processors implement the data processing and analysis method provided in the first aspect of the embodiment of the present invention.

[0042] According to another aspect of an embodiment of the present invention, a computer readable medium is provided.

[0043] The computer-readable medium according to an embodiment of the present invention stores a computer program thereon, and is characterized in that when the program is executed by a processor, the method for data processing and analysis provided by the first aspect of the embodiment of the present invention is implemented.

[0044] One embodiment of the above invention has the following advantages or beneficial effects: because of the technical means of predicting the duration of the user's next shopping, the technical problem of being unable to make an overall plan and control of the user's next shopping process within a limited time is overcome, thereby achieving the technical effect of predicting whether the user's next shopping is purposeful and improving the pertinence and accuracy of the product recommendation plan, and further provides effective parameters for other e-commerce-based optimization methods.

[0045] The further effects of the above-mentioned non-conventional optional manner will be described below in conjunction with the specific implementation manner. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] The accompanying drawings are used to better understand the present invention and do not constitute an improper limitation of the present invention.

[0047] Figure 1 is a schematic diagram of the main flow of the method for data processing and analysis according to an embodiment of the present invention;

[0048] Figure 2 is a schematic diagram of a database process according to an embodiment of the present invention;

[0049] Figure 3 is a schematic diagram of the overall process of the model according to an embodiment of the present invention;

[0050] Figure 4 is a schematic diagram of the process of another method for data processing and analysis according to an embodiment of the present invention;

[0051] Figure 5 is a schematic diagram of main modules of a device for data processing and analysis according to an embodiment of the present invention;

[0052] Figure 6 is an exemplary system architecture diagram to which embodiments of the present invention may be applied;

[0053] Figure 7 It is a schematic diagram of the structure of a computer system of a terminal device or a server suitable for implementing an embodiment of the present invention. DETAILED DESCRIPTION

[0054] The following is a description of exemplary embodiments of the present invention in conjunction with the accompanying drawings, including various details of the embodiments of the present invention to facilitate understanding, which should be considered as merely exemplary. Therefore, it should be recognized by those of ordinary skill in the art that various changes and modifications may be made to the embodiments described herein without departing from the scope and spirit of the present invention. Similarly, for clarity and conciseness, the description of well-known functions and structures is omitted in the following description.

[0055] According to one aspect of an embodiment of the present invention, a method for data processing and analysis is provided.

[0056] Figure 1 is a schematic diagram of the main flow of the method for data processing and analysis according to an embodiment of the present invention, such as Figure 1 As shown, the method for data processing and analysis according to an embodiment of the present invention includes steps S101, S102, S103, S104 and S105.

[0057] Step S101: construct a training database and a prediction database related to the user.

[0058] The database in this application can be constructed based on the background data of the e-commerce platform. Because for the e-commerce scenario, the computer can obtain the user and time-related behavior data from the background data and then construct a specific database according to different needs. When users shop on e-commerce websites, they are often accompanied by behaviors such as browsing, clicking, adding shopping carts, and at the same time, there are also user comment information, which are helpful for our modeling. In one embodiment, the background data of the e-commerce platform can be obtained through the mobile terminal used by the user, for example, a desktop computing device, a laptop computing device, an independent hardware device at least partially dedicated to an automatic assistant, a tablet computing device, a mobile phone computing device, a computing device of the user's vehicle (for example, a vehicle-mounted communication system, a vehicle-mounted entertainment system, a vehicle-mounted navigation system) or a wearable device of a user including a computing device (for example, a user's watch with a computing device, a user's glasses with a computing device, a virtual or augmented reality computing device). Additional and / or alternative client computing devices can be provided.

[0059] In one embodiment, the training database associated with the user is used to provide learning data for establishing a prediction model, and the prediction database associated with the user is used as the input of the trained model to predict the next shopping time or times of the user. "Shopping time" refers to the time a user stays on the online shopping platform, that is, the time from the user entering the online shopping platform to leaving the online shopping platform. "Predicted shopping time associated with the user" refers to the predicted time length value of the next time the user enters the online shopping platform to leave the online shopping platform predicted by the method provided in this application, which can use any time unit, such as minutes, seconds, etc. Therefore, the terms related to the concept of "shopping time" involved in this application should be interpreted accordingly in combination with the scenario on this basis. The introduction of the indicator "shopping time" can further derive more subdivided indicators, including but not limited to the time from the user browsing a specific product to placing an order, or the time interval between the end of this shopping and the start of the next shopping.

[0060] Step S102: extracting behavior feature recognition data related to the user from the training database.

[0061] For example, for e-commerce scenarios, we can obtain user and time-related behavior data. When users shop on e-commerce websites, they often browse, click, add to cart, and so on. At the same time, there is also user comment information, which is helpful for our modeling.

[0062] For our modeling, we first need to define an optimization goal, that is, we define the time period from the user entering the online shopping platform to leaving the online shopping platform as our optimization objective function. Obviously, our optimization objective function is a continuous value, so we can define the problem as a regression problem, so we can define our predicted shopping time y pre And the actual shopping time of our users ind The difference is taken as the loss of our model, which is the goal that our model needs to optimize, so we get the objective function This way we can further model based on this optimization objective function.

[0063] For example, our model is to predict the length of time a user will spend on their next shopping trip. We can understand that this is largely determined by the user's shopping habits. Some users prefer to browse online shopping platforms without much purpose. These users often spend a long time on each shopping trip. For some users who shop with a strong purpose, the time spent is often relatively short. Then we can further subdivide it. For users who shop with a purpose, users who are used to buying clothes on online shopping platforms and users who are used to buying mobile phones and other electronic products on online shopping platforms spend different amounts of time each time. Therefore, we can analyze that the length of time spent on each shopping trip by users of the same type is likely to be similar. Therefore, we can first consider using a clustering algorithm to cluster our user groups. Here, we use the k-means clustering algorithm to cluster our users. We cluster users by calculating the similarity between users. Here, we calculate the Euclidean distance. Here we set a hyperparameter K, which refers to how many categories we need to classify users into, and can adjust it later.

[0064] After we cluster users, we can extract corresponding aggregate features based on our clustering results, and we can also use the clustering results as the input of our model. Before introducing the feature extraction scheme, we first introduce the modeling database construction method for this modeling scheme. For the original database of our online shopping platform, we can get the original behavior data of users over a period of time, so our goal is to predict the user's future shopping time, time correlation, user granularity, and for our modeling this time, we do not need manual offline labeling. We can get the label for our model training by taking the user's shopping time online. For our data mining modeling, we need three databases, training set, validation set, and prediction set. The prediction set is our real online application scenario, so we can train our model through the training set, and verify and adjust our model on the validation set. When a certain effect is achieved, we can make online predictions.

[0065] The following is combined with Figure 2 , the division of offline training sets and validation sets is introduced. Figure 2 2 is a schematic diagram of a database process according to an embodiment of the present invention. The training database 210, the verification database 220 and the prediction database 230 obtained from the online user behavior database 240 can be data of three consecutive time units, for example, the behavior data of June, July and August related to the user. In order to establish a prediction model for the shopping time of the user, the extracted features 211 related to the user are obtained from the training database 210, and the extracted features 211 are fitted to obtain the optimization objective function 212. Optionally, the prediction model at this time can be verified by using the verification database 220, the extracted features 221 related to the user are obtained from the verification database 220, the extracted features 221 are fitted to obtain the optimization objective function 222, and the model is adjusted based on this. Finally, prediction can be made on the basis of the prediction database 230 based on the adjusted or unadjusted model, that is, the extracted features 231 related to the user are obtained from the prediction database 230, and the extracted features 231 are fitted to obtain the optimization objective function 232.

[0066] Optionally, the behavior feature recognition data includes one or more of the following: user basic features, user cluster group features, user product interaction features, and user text features.

[0067] For example, after the database is constructed, we can start to extract features. Here we can divide the features into several categories, user features, user cluster group features, and user product interaction features. First of all, for user features, we can get basic features, such as the user's age, gender, and purchasing power. Then we can analyze the business. First, we need to predict the length of the user's future shopping period, which must be strongly related to the user's historical shopping time. We can derive features based on this strongly related feature, so we can extract the length of the user's last shopping time, the average length of the user's historical shopping time, the user's historical total shopping time, and the variance of the user's historical shopping time. Similarly, we can also extract other features based on the interaction progress of time, such as the number of days between each shopping of the user, the day when the user last shopped, etc. Next is the user's cluster group feature. In the previous step, we have clustered the users, so we can extract some aggregate features for the clustered categories based on the results of our clustering of users, such as the historical average shopping time of each user category, etc. The simplest way is to do another layer of aggregation operation based on the extraction of user features mentioned in the previous step. In this way, we can extract rich aggregate features about user categories. Next is our user-product interaction feature. To extract this part of the feature, we need to interact with users and products, such as the number of products viewed by users each time they shop, which product category has the largest number of product categories viewed by users each time they shop, the top K product categories viewed by users, and the time it takes for users to browse, add to shopping carts, and place orders each time. At the same time, we can also do more detailed information, such as the time it took for users to browse products for the last time. And the time it took for users to browse products for the first time. Then we can combine the IDs of products that users have browsed, and use text conversion methods similar to tf-idf to perform corresponding conversions. We can regard the IDs of products that users have browsed as a text sequence, so we can use this method to learn the relationship between the IDs of products that users have browsed, and then provide richer user-product interaction features to help our modeling.

[0068] Step S103: Fitting the behavior feature recognition data to obtain an optimization objective function related to the user's shopping time.

[0069] Optionally, in the user-product interaction feature, a time-related ID code is performed on the product related to the user.

[0070] In one embodiment, for the product IDs that users have identified with behavioral features, this patent proposes a new encoding method. First, the purpose of our encoding is to process the various category features of the product ID into features that our model can learn. At the same time, because there are many categories, for an e-commerce platform, it contains hundreds of thousands or even millions of category features, so this category feature also contains very rich information. The following introduces an encoding method proposed by this patent in combination with this scenario. First, every time a user spends time on an e-commerce platform, he or she must have interacted with a certain product. So for a certain product X, assuming that the list of users who have browsed it is [u 1, u 2, u 3 ,…u n ], then the time each user spends on the product corresponds to the user list above [t 1, t 2, t 3 ,…t n ], then we define the value of the commodity as In this way, we convert the product ID of the category into a numerical value. Similarly, we can use this method to encode our product category ID and our product brand ID. This method not only captures our ID features that contain rich information, but also interacts with our duration, which is very beneficial to our modeling and prediction.

[0071] In one implementation, another important part of our modeling is the user's comment text data. The user's comment text is also an important aspect of characterizing the user. The user's satisfaction with the product and the user's satisfaction with the platform can indirectly affect our prediction of the duration of the user's next shopping trip. So how do we process these files? Here we propose a method of vectorizing these texts. By combining Tf-idf and Word2vec, for our original text, we first clean the special symbols and perform word segmentation. Then we perform Word2vec conversion and Tf-idf conversion on the segmented text. We vectorize the text by these two methods, and then we splice the vectorized results to get our final text.

[0072] Optionally, the user cluster group feature is a single-layer cluster feature or a multi-layer cluster feature related to the user.

[0073] After we have finished extracting the features, we can then use lightgbm to train our model, that is, to fit our features. As we mentioned before in the database partitioning, we train our model in the training set, and the model loss function is consistent with the evaluation function, that is, the mse loss function mentioned above. We can predict our model in the validation set and make corresponding parameter adjustments, because this model uses a tree model, such as the number of leaf nodes in the tree, etc. When our model meets the requirements for going online, we can predict the duration of our users' next shopping online, and then our platform can formulate more sophisticated marketing strategies based on the prediction results.

[0074] The purpose of the present invention is to provide a modeling solution for predicting the shopping duration of each user, fill the current gap in the exploration of modeling of user shopping duration, and contribute to the realization of smart shopping.

[0075] Step S104: establishing a first prediction model based on the optimization objective function.

[0076] Combination width Figure 3 Briefly describe the overall process of the model. Figure 3 It is a schematic diagram of the overall process of the model according to an embodiment of the present invention.

[0077] In one embodiment, user basic features 302, user cluster group features 303, user text features 304, and user-product interaction features 305 are extracted from the original data 301. Among them, the present application also provides a method for ID encoding the product features in the user-product interaction features 305 to obtain ID coding features 306, the main purpose of which is to represent the product as a time-related feature. This is because our optimization objective function is in units of time. This method makes the representation of the product strongly correlated with the optimization objective function, further enhancing the accuracy of the prediction results. Afterwards, the user basic features 302, user cluster group features 303, user text features 304, and user-product interaction features 305 and / or ID coding features 306 are fitted to obtain the total features 307. The total features 307 are used as the input of lightgbm308, and the results are used as the input of the validation set 309. Among them, lightgbm is a common machine learning algorithm.

[0078] This step may be as described in the above embodiment, which is merely exemplary and does not limit the steps.

[0079] Step S105: using the first prediction model to obtain the predicted shopping duration associated with the user based on the prediction database.

[0080] Optionally, using the first prediction model to obtain the predicted shopping time related to the user based on the prediction database further includes: constructing a verification database related to the user; using the verification database to verify and adjust the first prediction model to obtain a second prediction model; and using the second prediction model to obtain the predicted shopping time related to the user based on the prediction database.

[0081] In one embodiment, the step of "constructing a verification database related to the user" can be completed in step S101, that is, constructing a training database, a verification database and a prediction database related to the user. In one embodiment, the step of "using the verification database to verify and adjust the first prediction model to obtain a second prediction model" can be completed in step S104, or thereafter, for the purpose of using a verification database different from the training database to verify the accuracy of the first prediction model and fine-tune the first prediction model according to the verification result to obtain a second prediction model that is more accurate than the first prediction model. On the basis of obtaining the second prediction model, the predicted shopping duration related to the user is obtained based on the prediction database using the second prediction model. Exemplarily, the training database, verification database and prediction database related to the user can be the data of June, July and August acquired and stored by the e-commerce platform respectively. That is to say, in this example, the training database (June data) can be used for modeling, the verification database (July data) can be used to verify and adjust the established model, and finally the model can be used to predict the user's next shopping duration based on the prediction database (August data).

[0082] In e-commerce scenarios, how to use data-driven to provide more helpful information to our decision makers has been a hot topic in recent years. With the accumulation of big data, we have been able to achieve data-driven in many scenarios through big data plus machine learning technology. For example, store sales forecast. That is, we can use the store's past data and user behavior data to predict our store sales. For example, user purchase prediction, we can predict whether the user will make a purchase in the next day or month through the user's behavioral feature recognition data. For example, user click prediction, we can predict whether the user will click on a certain product.

[0083] In the context of data-driven development, the embodiments of the present invention explore the use of machine learning models to explore new prediction scenarios and predict the duration of a user's future shopping trip, thereby enabling us to predict in advance how long the user will stay on the online shopping platform next time, helping us to formulate more refined operation strategies.

[0084] The technical solution according to the embodiment of the present invention has the following advantages or beneficial effects: because the technical means of predicting the duration of the user's next shopping is adopted, the technical problem of being unable to make an overall plan and control of the user's next shopping process within a limited time is overcome, thereby achieving the technical effect of predicting whether the user's next shopping is purposeful and improving the pertinence and accuracy of the product recommendation plan, and further provides effective parameters for other e-commerce-based optimization methods.

[0085] According to another aspect of an embodiment of the present invention, a method for data processing and analysis is provided.

[0086] Figure 4 is a schematic diagram of the process of another method for data processing and analysis according to an embodiment of the present invention, such as Figure 4 As shown, the method for data processing and analysis according to an embodiment of the present invention includes steps S401, S402, S403, S404, S405 and S406.

[0087] Step S401: construct a training database, a verification database and a prediction database related to the user.

[0088] Step S402: extracting behavior feature recognition data related to the user from the training database.

[0089] Optionally, the behavior feature recognition data includes one or more of the following: user basic features, user cluster group features, user product interaction features, and user text features.

[0090] Optionally, in the user-product interaction feature, a time-related ID code is performed on the product related to the user.

[0091] Optionally, the user cluster group feature is a single-layer cluster feature or a multi-layer cluster feature related to the user.

[0092] Step S403: Fitting the behavior feature recognition data to obtain an optimization objective function related to the user's shopping time.

[0093] Step S404: establishing a first prediction model based on the optimization objective function.

[0094] Step S405: Use the verification database to verify and adjust the first prediction model to obtain a second prediction model.

[0095] Step S406: using the second prediction model to obtain the predicted shopping duration associated with the user based on the prediction database.

[0096] The technical solution according to the embodiment of the present invention has the following advantages or beneficial effects: because the technical means of predicting the duration of the user's next shopping is adopted, the technical problem of being unable to make an overall plan and control of the user's next shopping process within a limited time is overcome, thereby achieving the technical effect of predicting whether the user's next shopping is purposeful and improving the pertinence and accuracy of the product recommendation plan, and further provides effective parameters for other e-commerce-based optimization methods.

[0097] According to yet another aspect of an embodiment of the present invention, a device for data processing and analysis is provided.

[0098] Figure 5 is a schematic diagram of the main modules of the data processing and analysis device according to an embodiment of the present invention, such as Figure 5 As shown, the data processing and analysis device 500 according to an embodiment of the present invention includes a database construction module 501, a feature group extraction module 502, an optimization objective function fitting module 503, a model building module 504 and a shopping duration prediction module 505.

[0099] A database construction module 501 is used to construct a training database and a prediction database related to the user;

[0100] A feature group extraction module 502 is used to extract behavioral feature identification data related to the user from the training database;

[0101] An optimization objective function fitting module 503 is used to fit the behavior feature recognition data to obtain an optimization objective function related to the user's shopping time;

[0102] A model building module 504, configured to build a first prediction model based on the optimization objective function; and

[0103] The shopping duration prediction module 505 is used to obtain the predicted shopping duration related to the user based on the prediction database using the first prediction model.

[0104] Optionally, the behavior feature recognition data includes one or more of the following: user basic features, user cluster group features, user product interaction features, and user text features.

[0105] Optionally, the user cluster group feature is a single-layer cluster feature or a multi-layer cluster feature related to the user.

[0106] Optionally, the data processing and analysis device 500 according to the embodiment of the present invention further includes:

[0107] The ID encoding module 506 is used to perform time-related ID encoding on the commodities related to the user in the user-commodity interaction feature.

[0108] Optionally,

[0109] The database construction module 501 is also used to construct a verification database related to the user;

[0110] The model building module 504 is further configured to use the verification database to verify and adjust the first prediction model to obtain a second prediction model; and

[0111] The shopping duration prediction module 505 is further configured to use the second prediction model to obtain a predicted shopping duration associated with the user based on the prediction database.

[0112] The technical solution according to the embodiment of the present invention has the following advantages or beneficial effects: because the technical means of predicting the duration of the user's next shopping is adopted, the technical problem of being unable to make an overall plan and control of the user's next shopping process within a limited time is overcome, thereby achieving the technical effect of predicting whether the user's next shopping is purposeful and improving the pertinence and accuracy of the product recommendation plan, and further provides effective parameters for other e-commerce-based optimization methods.

[0113] Figure 6 An exemplary system architecture 600 is shown to which the data processing and analysis method or data processing and analysis device according to the embodiment of the present invention can be applied.

[0114] like Figure 6 As shown, system architecture 600 may include terminal devices 601, 602, 603, network 604 and server 605. Network 604 is used to provide a medium for communication links between terminal devices 601, 602, 603 and server 605. Network 604 may include various connection types, such as wired, wireless communication links or optical fiber cables, etc.

[0115] Users can use terminal devices 601, 602, and 603 to interact with server 605 through network 604 to receive or send messages, etc. Various communication client applications can be installed on terminal devices 601, 602, and 603, such as shopping applications, web browser applications, search applications, instant messaging tools, email clients, social platform software, etc. (only examples).

[0116] The terminal devices 601 , 602 , and 603 may be various electronic devices having a display screen and supporting web browsing, including but not limited to smart phones, tablet computers, laptop computers, and desktop computers, etc.

[0117] The server 605 may be a server that provides various services, such as a backend management server (only an example) that provides support for shopping websites browsed by users using the terminal devices 601, 602, and 603. The backend management server may analyze and process the received data such as product information query requests, and feed back the processing results (such as target push information, product information - only an example) to the terminal device.

[0118] It should be noted that the data processing and analysis method provided in the embodiment of the present invention is generally executed by the server 605 , and accordingly, the data processing and analysis device is generally disposed in the server 605 .

[0119] It should be understood that Figure 6 The number of terminal devices, networks and servers in the embodiment is only for illustration. Any number of terminal devices, networks and servers may be provided according to the implementation requirements.

[0120] Reference below Figure 7 , which shows a schematic diagram of the structure of a computer system 700 of a terminal device suitable for implementing an embodiment of the present invention. Figure 7 The terminal device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present invention.

[0121] like Figure 7 As shown, the computer system 700 includes a central processing unit (CPU) 701, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 702 or a program loaded from a storage part 708 into a random access memory (RAM) 703. In the RAM 703, various programs and data required for the operation of the system 700 are also stored. The CPU 701, the ROM 702, and the RAM 703 are connected to each other via a bus 704. An input / output (I / O) interface 705 is also connected to the bus 704.

[0122] The following components are connected to the I / O interface 705: an input section 706 including a keyboard, a mouse, etc.; an output section 707 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker, etc.; a storage section 708 including a hard disk, etc.; and a communication section 709 including a network interface card such as a LAN card, a modem, etc. The communication section 709 performs communication processing via a network such as the Internet. A drive 710 is also connected to the I / O interface 705 as needed. A removable medium 711, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 710 as needed, so that a computer program read therefrom is installed into the storage section 708 as needed.

[0123] In particular, according to the embodiments disclosed in the present invention, the process described above with reference to the flowchart can be implemented as a computer software program. For example, the embodiments disclosed in the present invention include a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network through the communication part 709, and / or installed from the removable medium 711. When the computer program is executed by the central processing unit (CPU) 701, the above-mentioned functions defined in the system of the present invention are executed.

[0124] It should be noted that the computer-readable medium shown in the present invention may be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present invention, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, device or device. In the present invention, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, which carries a computer-readable program code. This propagated data signal may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. Computer-readable signal media may also be any computer-readable medium other than computer-readable storage media, which may send, propagate or transmit a program for use by or in conjunction with an instruction execution system, apparatus or device. The program code contained on the computer-readable medium may be transmitted using any appropriate medium, including but not limited to: wireless, wire, optical cable, RF, etc., or any suitable combination of the above.

[0125] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present invention. In this regard, each box in the flow chart or block diagram can represent a module, a program segment, or a part of a code, and the above-mentioned module, program segment, or a part of a code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram or flow chart, and the combination of the boxes in the block diagram or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0126] The modules involved in the embodiments of the present invention may be implemented by software or hardware. The modules described may also be set in a processor, for example, it may be described as: a processor including a database construction module, a feature group extraction module, an optimization objective function fitting module, a model building module and a shopping duration prediction module. The names of these modules do not constitute a limitation on the modules themselves in some cases, for example, the model building module may also be described as "a module for establishing a first prediction model based on the optimization objective function".

[0127] As another aspect, the present invention also provides a computer-readable medium, which may be included in the device described in the above embodiment; or it may exist independently without being assembled into the device. The above computer-readable medium carries one or more programs, and when the above one or more programs are executed by a device, the device includes: constructing a training database and a prediction database related to the user; extracting behavior feature identification data related to the user from the training database; fitting the behavior feature identification data to obtain an optimization objective function related to the user's shopping time; establishing a first prediction model based on the optimization objective function; and using the first prediction model to obtain a predicted shopping time related to the user based on the prediction database.

[0128] The technical solution according to the embodiment of the present invention has the following advantages or beneficial effects: because the technical means of predicting the duration of the user's next shopping is adopted, the technical problem of being unable to make an overall plan and control of the user's next shopping process within a limited time is overcome, thereby achieving the technical effect of predicting whether the user's next shopping is purposeful and improving the pertinence and accuracy of the product recommendation plan, and further provides effective parameters for other e-commerce-based optimization methods.

[0129] The above specific implementations do not constitute a limitation on the protection scope of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions may occur depending on design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for data processing and analysis, characterized in that: include: Based on the user's behavior data and comment information on the online shopping platform, construct a user-related training database and prediction database; Extracting behavioral feature identification data related to the user from the training database; The behavior feature identification data includes: user basic features, user cluster group features, user commodity interaction features and user text features; wherein, in the user commodity interaction features, the commodities related to the user are coded with time-related IDs; the user text features include: comment text data; Use lightgbm to fit the behavior feature recognition data to obtain an optimization objective function related to the user's shopping time; the optimization objective function is: the time period from when the user enters the online shopping platform to when he leaves the online shopping platform; Establishing a first prediction model based on the optimization objective function; and The first prediction model is used to obtain a predicted shopping duration associated with the user based on the prediction database.

2. The method according to claim 1, characterized in that in, The user cluster group feature is a single-layer cluster feature or a multi-layer cluster feature related to the user.

3. The method according to claim 1, characterized in that Using the first prediction model to obtain the predicted shopping duration related to the user based on the prediction database further includes: Construct a user-related authentication database; Using the verification database to verify and adjust the first prediction model to obtain a second prediction model; and The second prediction model is used to obtain a predicted shopping duration related to the user based on the prediction database.

4. A data processing and analysis device, characterized in that: include: A database construction module is used to construct a training database and a prediction database related to users based on their behavior data and comment information on the online shopping platform; A feature group extraction module, used to extract behavioral feature identification data related to the user from the training database; The behavior feature identification data includes: user basic features, user cluster group features, user commodity interaction features and user text features; wherein, in the user commodity interaction features, the commodities related to the user are coded with time-related IDs; the user text features include: comment text data; An optimization objective function fitting module is used to use lightgbm to fit the behavior feature recognition data to obtain an optimization objective function related to the user's shopping time; the optimization objective function is: the time period from the user entering the online shopping platform to leaving the online shopping platform; A model building module, used to build a first prediction model based on the optimization objective function; and A shopping duration prediction module is used to obtain a predicted shopping duration related to the user based on the prediction database using the first prediction model.

5. The device according to claim 4, characterized in that in, The user cluster group feature is a single-layer cluster feature or a multi-layer cluster feature related to the user.

6. The device according to claim 4, characterized in that in, The database construction module is also used to construct a verification database related to the user; The model building module is further used to verify and adjust the first prediction model using the verification database to obtain a second prediction model; and The shopping duration prediction module is further used to obtain the predicted shopping duration related to the user based on the prediction database using the second prediction model.

7. An electronic device for data processing and analysis, characterized in that: include: one or more processors; a storage device for storing one or more programs, When the one or more programs are executed by the one or more processors, the one or more processors implement the method according to any one of claims 1 to 3.

8. A computer readable medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the method according to any one of claims 1 to 3 is implemented.

Images