Short message reach analysis method and device, computer device and readable storage medium

Abstract

The application discloses a short message reach analysis method based on a data model. The short message reach analysis method based on the data model comprises the following steps: obtaining usage data of a user using an application program, the usage data being related to a short message service and / or an order placing behavior; analyzing, based on the usage data and the data model, an analysis result of whether the short message reach is used; verifying the analysis result based on a preset verification model to obtain a verification result; and modifying the data model based on the verification result. The application also discloses a short message reach analysis device based on the data model, a computer device and a computer readable storage medium. The data model obtains the analysis result of whether the short message reach is used based on the usage data of the user using the application program, the analysis result is further verified by the verification model, and the data model is further optimized according to the verification result, so that the objectivity and accuracy of the analysis result analyzed by the data model are higher.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data analysis, in particular to a short message reach analysis method based on a data model, a short message reach analysis device based on a data model, a computer device and a computer readable storage medium. BACKGROUND

[0002] In recent years, with the rapid development of the industry and the progress of technology, there are various ways to reach users through messages, and the efficiency and conversion rate of message reach are increasingly concerned by enterprises and consumers. The advantage of short message reach is that it can directly reach users without opening the application (app), but the disadvantage is that it is relatively more expensive than other reach methods and can easily disturb customers. Therefore, how to improve the efficiency of message reach, effectively reach, and reduce user complaints is a problem to be solved. SUMMARY

[0003] To solve at least one of the technical problems in the background art, the embodiments of the present application provide a short message reach analysis method based on a data model, a short message reach analysis device based on a data model, a computer device and a computer readable storage medium.

[0004] The short message reach analysis method based on a data model of the embodiments of the present application comprises:

[0005] Obtaining usage data of users using an application, the usage data being related to short message services and / or order placement behaviors;

[0006] Based on the usage data and the data model, an analysis result of whether to use short message reach is obtained;

[0007] Based on a preset verification model, the analysis result is verified to obtain a verification result; and

[0008] Based on the verification result, the data model is modified.

[0009] In some embodiments, the obtaining of the usage data of the users using the application comprises:

[0010] Obtaining a user category of each user as first data according to a response rate of the user after receiving a short message service;

[0011] Obtaining second data of each user according to operation data of the user after receiving a short message service for the short message service;

[0012] Obtaining third data of each user according to usage habit data of the user using the application; and

[0013] Obtaining fourth data of each user according to data of the user using the application to place an order.

[0014] In some embodiments, the data model is a LR model, and the first data, the second data, the third data and the fourth data are feature inputs of the LR model.

[0015] In some embodiments, the modifying the data model based on the verification result comprises:

[0016] modifying a weight ratio of the first data, the second data, the third data and the fourth data in the data model based on the verification result; and / or

[0017] modifying a bias in the data model based on the verification result.

[0018] In some embodiments, the usage habit data comprises at least one or more of a time interval from when the user starts the application to when the first order is created, a time interval from when the user uses the application to when the first order is created, and a time interval from when the user estimates the price to when the order is placed.

[0019] In some embodiments, the analysis result comprises one of selecting the SMS touch or selecting the push touch, and the verifying the analysis result based on the preset verification model and obtaining a verification result comprises:

[0020] when the analysis result is selecting the push touch, analyzing the order data under the push touch based on the verification model; and

[0021] comparing the order data with a preset threshold to obtain the verification result.

[0022] In some embodiments, the analysis result comprises one of selecting the SMS touch or selecting the push touch, and the verifying the analysis result based on the preset verification model and obtaining a verification result comprises:

[0023] when the analysis result is selecting the push touch, respectively obtaining the order data of the SMS touch and the order data of the push touch based on an ABTest method; and

[0024] comparing the order data of the SMS touch and the order data of the push touch to obtain the verification result.

[0025] The SMS touch analysis device based on a data model according to the embodiments of the present application comprises:

[0026] an obtaining module, configured to obtain usage data of a user using an application, the usage data being related to an SMS service and / or an order behavior;

[0027] an analysis module configured to analyze, based on the use data and the data model, an analysis result of whether to use short message service (SMS) reach.

[0028] a verification module configured to verify, based on a preset verification model, the analysis result and obtain a verification result; and

[0029] a modification module configured to modify, based on the verification result, the data model.

[0030] The computer device of the embodiments of the present application includes one or more processors, a memory, and one or more computer programs, wherein the one or more computer programs are stored in the memory and configured to be executed by the one or more processors, and the one or more computer programs are configured to perform the data model-based SMS reach analysis method described in any of the embodiments of the present application.

[0031] The non-volatile computer readable storage medium of the embodiments of the present application stores a computer program, which, when executed by one or more processors, causes the processors to perform the data model-based SMS reach analysis method described in any of the embodiments of the present application.

[0032] In the data model-based SMS reach analysis method, the data model-based SMS reach analysis device, the computer device, and the computer readable storage medium of the embodiments of the present application, the data model is based on use data of a user using an application program to obtain an analysis result of whether to use SMS reach. The analysis result is further verified by a verification model, and the data model is further optimized according to the verification result, so that the objectivity and accuracy of the analysis result analyzed by the data model are higher, and the selection of the reach method according to the analysis result is beneficial to improve the conversion efficiency of reach and reduce the complaint rate.

[0033] Additional aspects and advantages of the embodiments of the present application will be in part apparent and in part pointed out hereinafter in the description of the embodiments of the present application. BRIEF DESCRIPTION OF DRAWINGS

[0034] The above and / or additional aspects and advantages of the present application can become apparent and more readily appreciated from the following description of the embodiments, taken in conjunction with the accompanying drawings in which:

[0035] Figure 1 A flowchart of the data model-based SMS reach analysis method of some embodiments of the present application;

[0036] Figure 2 A flowchart of the data model-based SMS reach analysis method of some embodiments of the present application;

[0037] Figure 3 Flowchart of a data model-based SMS reach analysis method according to some embodiments of the present application;

[0038] Figure 4 Flowchart of a data model-based SMS reach analysis method according to some embodiments of the present application;

[0039] Figure 5 Flowchart of a data model-based SMS reach analysis method according to some embodiments of the present application;

[0040] Figure 6 Module diagram of a data model-based SMS reach analysis apparatus according to some embodiments of the present application;

[0041] Figure 7 Schematic diagram of a computer readable storage medium in communication with a processor according to some embodiments of the present application;

[0042] Figure 8 Schematic diagram of a computer device according to some embodiments of the present application. DETAILED DESCRIPTION

[0043] Embodiments of the present application are described in detail below with reference to the attached drawing figures, wherein the same or like designations indicate the same or like elements or elements having the same or similar functionality. The embodiments described below are merely exemplary for the purpose of explanation and are not to be understood as limiting the present application.

[0044] In recent years, with the rapid development of the industry and the progress of technology, there are various ways to reach users through messages, and the efficiency and conversion rate of message reach are increasingly concerned by enterprises and consumers. How to improve the efficiency of message reach, become effective reach, and reduce user complaints is a problem that needs to be solved. In the field of message reach, push push (relying on the app on the phone to reach messages), SMS push, email push, and voice notification (through phone voice to continue message reach) are common ways of reach. Compared with the three push methods of SMS push, email push, and voice notification, push push has lower cost, so it is necessary to use push push as much as possible to reach messages, but not all cases can use push push to reach messages. Push push can be shielded by users, and the sensitivity of users to push push is not as high as that of SMS push. The advantage of SMS reach is that it can directly reach the client without opening the app, but the disadvantage is that the cost is relatively high compared with other channels, and it is easy to harass the client.

[0045] The short message touch analysis method based on the data model of the embodiments of the present application can reasonably select whether to use the short message touch mode, shield part of the low-frequency users without conversion effect, and not send short messages to reduce the cost and complaint rate while ensuring the message touch conversion rate.

[0046] Please refer to Figure 1 , Figure 1 The flowchart of the short message touch analysis method based on the data model of some embodiments of the present application is shown in the figure. The short message touch analysis method comprises the following steps:

[0047] 01: Obtain the use data of the user using the application program, and the use data is related to the short message service and / or order behavior;

[0048] 02: Based on the use data and the data model, analyze the analysis result of whether to use the short message touch;

[0049] 03: Based on the preset verification model, verify the analysis result and obtain the verification result; and

[0050] 04: Based on the verification result, modify the data model.

[0051] In the short message touch analysis method based on the data model of the embodiments of the present application, the data model obtains the analysis result of whether to use the short message touch based on the use data of the user using the application program. The analysis result is further verified by the verification model, and the data model is further optimized according to the verification result, so that the objectivity and accuracy of the analysis result analyzed by the data model are higher, and the selection of the touch mode according to the analysis result is beneficial to improve the conversion efficiency of the touch and reduce the complaint rate.

[0052] Specifically, in the implementation of step 01, the use data of the user using the application program is obtained, wherein the use data is related to the short message service and / or order behavior. Therefore, the use data can be used to reflect the user's operation reaction to the short message service, or can be used to reflect the user's order operation habit, etc. Therefore, the use data can be used as an input parameter of the data model to further analyze whether it is suitable to perform short message touch on the user.

[0053] When step 02 is implemented, based on the usage data and the data model, an analysis result of whether to use short message reach is obtained, the data model is used to analyze whether the user is suitable for using short message reach, if not suitable for short message reach, whether it is suitable for other reach modes, such as push reach, email reach, telephone reach, etc. After obtaining the analysis result, step 03 is implemented, the analysis result is verified by using a verification model and a verification result is obtained, the verification model is different from the data model, and the analysis result is verified by the verification model to further ensure that the finally used reach mode is more reasonable. Further, when step 04 is implemented, based on the verification result, the data model is modified, when the verification result is that the analysis result is not accurate enough, the data model is further modified to continuously improve the data model, when the verification result is that the analysis result is accurate enough, the data model can not be modified.

[0054] Please refer to Figure 2 , Figure 2 is a flowchart of a data model-based short message reach analysis method of some embodiments of the application, in some embodiments, step 01: obtaining usage data of users using application programs, including steps:

[0055] 011: obtaining the user category of each user as first data according to the response rate of the user after receiving the short message service;

[0056] 012: obtaining the second data of each user according to the operation data of the user to the short message service after receiving the short message service;

[0057] 013: obtaining the third data of each user according to the usage habit data of the user using the application program; and

[0058] 014: obtaining the fourth data of each user according to the data of the user using the application program to place an order.

[0059] Specifically, when step 011 is implemented, the user category of each user is obtained as first data according to the response rate of the user after receiving the short message service, more specifically, in one example, the user can be divided into seven categories according to the response rate of the user to the received short message service, which are: receiving short message in the first 15 days and not using app in the first 15 days; receiving short message in the first 15 days and using app in the first 15 days; not receiving short message in the first 15 days and using app in the first 15 days; not receiving short message in the first 15 days and not using app in the first 15 days, registered within 30 days; not receiving short message in the first 15 days and not using app in the first 15 days, registered within 30-60 days; not receiving short message in the first 15 days and not using app in the first 15 days, registered within 60-90 days; not receiving short message in the first 15 days and not using app in the first 15 days, registered more than 90 days, and the first data is the classification of the user.

[0060] When step 012 is implemented, the second data of each user is obtained according to the operation data of the short message service after the user receives the short message service, and specifically, in an example, the second data includes:

[0061] Short message reply & short message blacklist: after the batch of short message tasks are sent out, how many users do not want to receive them, and how many users have a negative attitude towards the batch of short messages, in order to improve user experience, these users can not continue to send short messages;

[0062] Short message scenario & short message content: because short message operators use different short message scenarios, but there are similar short message contents, operators check the same user group, resulting in repeated receipt of similar short messages by users, leading to a decline in user experience of short messages, and thus affecting the click rate of short messages;

[0063] Short message short chain click & app opening: how many users see the short message and are willing to open the short link on the short message, or even open the app after seeing the short message and use it, indicating that these users have a positive attitude towards the short message and are worth continuing to send;

[0064] App opening & ordering: receiving a short message and opening an app is a funnel conversion, and from opening an app to the ordering process, it is also a funnel conversion. Looking at the whole link, how many users open the app and complete the ordering operation after receiving the short message, which is the final revenue effect of the short message task;

[0065] Sending in the previous x days but not opening the app or short chain: these users do not reject short message sending, because they do not unsubscribe to the short message, but they are low in sensitivity to the short message, and may not have seen the short message content at all, belonging to a low use and low conversion rate group, which needs to be further subdivided and further accurately screened.

[0066] When step 013 is implemented, the third data of each user is obtained according to the user's use habit data of the application program, and in an example, the third data includes:

[0067] Time interval from starting the app to creating the first order: the order base table takes the first order of the user on the same day; the starting app record selects the nearest time from the order, calculates the time interval from starting the app to ordering, analyzes the user's heart journey from using the app to ordering, and analyzes what behavior operation is done in the time interval from starting to ordering, whether the app is hung in the background after starting, and whether the app is really used, what time period starts to use the app, why the background app is suddenly awakened, the app is used and the ordering behavior is generated, in order to facilitate analysis and selection of the timing of push.

[0068] □ Time interval from app use to first order creation: The order base table takes the first order of the user on the same day, and the app record takes the first time of use on the same day, and the time interval from app use to order is calculated. Analyze how long it takes for users to generate orders from app use, and what the time spent in the middle is mainly spent on. Is it comparing prices, or selecting goods, etc. Different user behaviors can be targeted by different push, such as coupon push for price comparison users, and product recommendation push for users selecting goods. Formulate the user's corresponding user portrait label.

[0069] Time interval from valuation to order: The user originally intended to place an order, but due to price problems, no order behavior was generated. You can view how many times the user generated valuation behavior, and finally generated order behavior. Establish the corresponding portrait label for this user, and push the coupon during the valuation period to promote the generation of order behavior.

[0070] When implementing step 014, according to the data of the user using the application program to place an order, the fourth data of each user is obtained. In one example, the fourth data can include the number of received messages, the number of users receiving messages, the proportion of the crowd, the number of users placing orders within one hour, the proportion of orders placed within one hour, the number of users placing orders on the same day, and the proportion of orders placed on the same day.

[0071] Further, in one example, the data model is an LR model, and the first data, the second data, the third data, and the fourth data are feature inputs of the LR model. One expression of the data model is:

[0072]

[0073] The unified form is

[0074] Where y is the predicted parameter (used to obtain the analysis result, i.e. push reach, short message reach, etc.); w is the model parameter, i.e. the weight ratio corresponding to the first data, the second data, the third data, and the fourth data; x is the feature input, i.e. the first data, the second data, the third data, and the fourth data; b is the bias, which is a parameter that needs to be adjusted slightly when modifying the data model; t is the subscript, representing the parameter.

[0075] In addition, make the following LR assumptions: First assumption: the data follows Bernoulli distribution (any model has assumptions, such as data classification features, so that the model can be classified); Second assumption: the probability of using data samples is a Sigmoid function, and the second assumption is used to construct the LR equation.

[0076] The logistic regression model equation is composed of the linear regression model equation and the Sigmoid function, that is, f(x) = w t x and Combined, the logistic regression model equation is obtained:

[0077]

[0078] Again, the logistic regression model equation is operated:

[0079] (1) The logistic function: for multivariate logistic regression, the following formula can be used to classify:

[0080]

[0081] Let:

[0082] For the training data set, the feature data x = {x1, x2, …, xm} and the corresponding classification data y = {y1, y2, …, ym}. The most typical method to construct the logistic regression model f(θ) is to apply maximum likelihood estimation. First, for a single sample, its posterior probability is: P(y|x, θ) = (hθ(x)) y (1-hθ(x)) 1-y Where y = 1 (or 0), then the maximum likelihood function is:

[0083]

[0084] The log likelihood is:

[0085]

[0086] (2) Gradient descent: optimize the parameter weight From section (1), to solve the logistic regression model f(θ), it is equivalent to:

[0087] θ * = arg min(l(θ))

[0088] Use gradient descent method:

[0089]

[0090] Thus, iterate θ to convergence: Get The final formula:

[0091]

[0092] According to the analysis result y, an analysis result of whether to use the short message touch is obtained. For example, in one example, when y is between 0 and 0.5, the analysis result is that the short message touch is suitable, and when y is between 0.5 and 1, the analysis result is that other touch methods such as push touch are suitable.

[0093] Further, please refer to Figure 3 , Figure 3 FIG. 1 is a flowchart of a data model-based short message touch analysis method according to some embodiments of the present application. In some embodiments, step 04: modifying the data model based on the verification result, includes the following steps:

[0094] 041: modifying the weight proportions of the first data, the second data, the third data and the fourth data in the data model based on the verification result; and / or

[0095] 042: modifying the bias in the data model based on the verification result.

[0096] By modifying the weight proportions and the bias of the LR model, the data model is further improved, so that the analysis result obtained by the data model is more objective and accurate.

[0097] Please refer to Figure 4 , Figure 4 FIG. 1 is a flowchart of a data model-based short message touch analysis method according to some embodiments of the present application. In some embodiments, the analysis result includes selecting one of the short message touch or the push touch, and step 03: verifying the analysis result based on a preset verification model to obtain a verification result, includes the following steps:

[0098] 031: when the analysis result is to select the push touch, analyzing the order data under the push touch based on the verification model; and

[0099] 032: comparing the order data with a preset threshold to obtain the verification result.

[0100] Specifically, when the analysis result for a certain user is to select the push touch, the user needs to be verified by the verification model, that is, to analyze the order data under the push touch, and to compare the order data with the preset threshold. When the order data is less than the preset threshold, it indicates that the effect of using the push touch for this user is not good, and the analysis result is not accurate enough, so the data model needs to be modified.

[0101] Specifically, when analyzing the order data under push reach, the following data can be analyzed: push magnitude, including how many times the minimum one-day amount is, how many times the maximum one-day amount is, whether there is a periodic rule of the amount; distribution of the time point of push; distribution of the page corresponding to the click of the push; and the top 20 push titles; under the algorithm strategy grouping, the people who receive the push can be divided into 4 categories: 1st category of users: sent push within the first 15 days, and never opened the app within the first 15 days (push invalid users), 2nd category of users: sent push within the first 15 days, and opened the app within the first 15 days (push valid users), 3rd category of users: never sent push within the first 15 days, and opened the app within the first 15 days (spontaneous opening users), and 4th category of users: never sent push within the first 15 days, and never opened the app within the first 15 days (new users); the push received by the newly registered users and the order amount per month are compared for analysis.

[0102] Referring to Figure 5 , Figure 5 The flowchart of the data model-based SMS reach analysis method of some embodiments of the present application is shown in the figure. In some embodiments, the analysis result includes selecting one of SMS reach or push reach, step 03: verifying the analysis result based on a preset verification model to obtain a verification result, including steps of:

[0103] 033: When the analysis result is to select push reach, the order data of SMS reach and the order data of push reach are obtained based on the ABTest method; and

[0104] 034: Comparing the order data of SMS reach and the order data of push reach to obtain the verification result.

[0105] AbTest can be used to verify the effect of the data model, and to measure the good and bad effects of the data model. Specifically, the same population is extracted, and push reach and SMS reach are used respectively to see how the conversion effect is. Different user groups can also be extracted. Originally, SMS reach was used, and then push reach was used to see how the effects of the two are. Comparing the order data of SMS reach and the order data of push reach to obtain the verification result.

[0106] Referring to Figure 6 , Figure 6For the schematic diagram of the module of the data model based SMS touch analysis device 10 of some embodiments of the present application, the SMS touch analysis device 10 comprises an acquisition module 11, an analysis module 12, a verification module 13 and a modification module 14. The acquisition module 11 can be used to implement step 01, i.e. the acquisition module 11 can be used to acquire the usage data of the user using the application program, and the usage data is related to the SMS service and / or the order placement behavior. The analysis module 12 can be used to implement step 02, i.e. the analysis module 12 can be used to analyze whether the SMS touch is used based on the usage data and the data model to obtain an analysis result. The verification module 13 can be used to implement step 03, i.e. the verification module 13 can be used to verify the analysis result based on the preset verification model to obtain a verification result. The modification module 14 can be used to implement step 04, i.e. the modification module 14 can be used to modify the data model based on the verification result.

[0107] Please continue to refer to Figure 6 In some embodiments, the acquisition module 11 can be used to implement steps 011, 012, 013 and 014, i.e. the acquisition module 11 can be used to acquire the user category of each user as the first data according to the response rate of the user after receiving the SMS service; acquire the second data of each user according to the operation data of the user for the SMS service after receiving the SMS service; acquire the third data of each user according to the usage habit data of the user using the application program; and acquire the fourth data of each user according to the data of the user using the application program to place an order.

[0108] Please continue to refer to Figure 6 In some embodiments, the modification module 14 can be used to implement steps 041 and 042, i.e. the modification module 14 can be used to modify the weight proportion of the first data, the second data, the third data and the fourth data in the data model based on the verification result; and / or modify the bias amount in the data model based on the verification result.

[0109] Please continue to refer to Figure 6 In some embodiments, the verification module 13 can be used to implement steps 031 and 032, i.e. the verification module 13 can be used to analyze the order placement data under the push touch based on the verification model when the analysis result is to select the push touch; and compare the order placement data with the preset threshold to obtain the verification result.

[0110] Please continue to refer to Figure 6 In some embodiments, the verification module 13 can be used to implement steps 033 and 034, i.e. the verification module 13 can be used to acquire the order placement data of the SMS touch and the order placement data of the push touch based on the ABTest method respectively when the analysis result is to select the push touch; and compare the order placement data of the SMS touch and the order placement data of the push touch to obtain the verification result.

[0111] It should be noted that the implementation details and effects of the short message reach analysis device 10 when implementing the short message reach analysis method of any embodiment of the present application can refer to the description of the short message reach analysis method above, and will not be described here.

[0112] In addition, referring to Figure 7 The embodiment of the present application provides a computer readable storage medium, and a computer program is stored on the computer readable storage medium. The computer program is executed by a processor to implement the short message reach analysis method of any of the above embodiments. The computer readable storage medium includes but is not limited to any type of disk (including floppy disk, hard disk, optical disk, CD-ROM, and magneto-optical disk), ROM (Read-Only Memory), RAM (Random Access Memory), EPROM (Erasable Programmable Read-Only Memory), EEPROM (Electrically Erasable Programmable Read-Only Memory), flash memory, magnetic card or optical card. That is, the storage device includes any medium that stores or transmits information in a form that can be read by a device (for example, a computer, a mobile phone), which can be a read-only memory, a magnetic disk or an optical disk, etc.

[0113] The content of the method embodiment of the present application is applicable to the storage medium embodiment of the present application. The storage medium embodiment specifically implements the same functions as the above method embodiments, and achieves the same beneficial effects as the above method embodiments. For details, please refer to the description in the method embodiment, which will not be described here.

[0114] In addition, referring to Figure 8 The embodiment of the present application further provides a computer device. The computer device of the embodiment can be a server, a personal computer, a network device, etc. The computer device includes one or more processors, a memory, and one or more computer programs. The one or more computer programs are stored in the memory and configured to be executed by the one or more processors. The one or more computer programs are configured to execute the short message reach analysis method of any of the above embodiments.

[0115] In the description of the specification, the description of the terms "one embodiment", "some embodiments", "exemplary embodiment", "example", "specific example" or "some examples" etc. means that the specific features, structures, materials or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present application. In the specification, the exemplary description of the above terms does not necessarily refer to the same embodiment or example. Also, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in an appropriate manner. In addition, the person skilled in the art can combine and combine the different embodiments or examples described in the specification and the features of the different embodiments or examples without contradiction.

[0116] Any process or method descriptions or descriptions of the flow diagrams in the specification or otherwise described herein can be understood as representing the steps of the code modules, segments, or portions of the respective code that include one or more executable instructions for performing specific logical functions or steps in the process. The scope of preferred embodiments of the present application includes the additional implementation in which the functions can be performed in an order different from the order shown or discussed, including functions performed in substantially simultaneous, or in reverse order, as appropriate, depending on the functions involved, as would be understood by a person skilled in the art of the embodiments of the present application.

[0117] Although the embodiments of the present application have been shown and described above, it is understood that the above-described embodiments are exemplary and should not be construed as limiting the present application, and those skilled in the art can make changes, modifications, replacements and variations to the above-described embodiments within the scope of the present application.

Claims

1. A data model based SMS reach analysis method, characterized in that, The short message reach analysis method comprises: obtaining usage data of users using an application, the usage data being related to short message service or order placement behavior, or the usage data being related to short message service and order placement behavior; based on the usage data and the data model, analyzing to obtain an analysis result of whether to use short message reach; based on a preset verification model, verifying the analysis result and obtaining a verification result; and based on the verification result, modifying the data model; the obtaining of the usage data of the users using the application comprises: obtaining a user category of each user as first data according to a response rate of the user after receiving short message service; obtaining second data of each user according to operation data of the user after receiving short message service for the short message service; obtaining third data of each user according to usage habit data of the user using the application; and obtaining fourth data of each user according to data of the user using the application to place an order; the first data, the second data, the third data and the fourth data are feature inputs of the data model.

2. The short message touch point analysis method of claim 1, wherein, The data model is an LR model.

3. The short message touch point analysis method of claim 2, wherein, The modifying of the data model based on the verification result comprises: based on the verification result, modifying a weight ratio of first data, second data, third data and fourth data in the data model; and / or based on the verification result, modifying a bias in the data model.

4. The short message touch point analysis method of claim 1, wherein, The usage habit data comprises at least one or more of a time interval from starting the application to creating a first order, a time interval from using the application to creating a first order, and a time interval from estimating a price to placing an order.

5. The short message touch point analysis method of claim 1, wherein, The analysis result comprises one of selecting short message reach or selecting push reach, and the verifying of the analysis result based on the preset verification model and obtaining a verification result comprises: when the analysis result is selecting push reach, analyzing order placement data under push reach based on the verification model; and comparing the order placement data with a preset threshold to obtain the verification result.

6. The short message touch point analysis method of claim 1, wherein, The analysis result comprises one of selecting short message reach or selecting push reach, and the verifying of the analysis result based on the preset verification model and obtaining a verification result comprises: when the analysis result is selecting push reach, respectively obtaining short message reach order placement data and push reach order placement data based on an ABTest method; and comparing the short message reach order placement data and the push reach order placement data to obtain the verification result.

7. A data model based SMS reach analysis apparatus, characterized by, The short message reach analysis device comprises: an obtaining module, configured to obtain usage data of users using an application, the usage data being related to short message service or order placement behavior, or the usage data being related to short message service and order placement behavior; an analyzing module, configured to analyze, based on the usage data and the data model, to obtain an analysis result of whether to use short message reach; a verifying module, configured to verify, based on a preset verification model, the analysis result and obtain a verification result; and a modifying module, configured to modify, based on the verification result, the data model. The obtaining the usage data of the user using the application program comprises: obtaining a user category of each user as first data according to a response rate of the user after receiving a short message service; obtaining second data of each user according to operation data of the user for the short message service after receiving the short message service; obtaining third data of each user according to usage habit data of the user using the application program; and obtaining fourth data of each user according to data of the user using the application program to place an order. The first data, the second data, the third data and the fourth data are feature inputs of the data model.

8. A computer device, comprising: comprise: one or more processors; a memory; and one or more computer programs, wherein the one or more computer programs are stored in the memory and configured to be executed by the one or more processors, and the one or more computer programs are configured to execute the method for analyzing short message reach according to the data model in any one of claims 1 to 6. 9.A non-transitory computer readable storage medium storing a computer program, wherein the computer program comprises instructions that, when executed by a processor, cause the processor to perform the method according to any one of claims 1 to 8. When the computer programs are executed by one or more processors, the processors execute the method for analyzing short message reach according to the data model in any one of claims 1 to 6.

Images