An internet user portrait generation method based on big data

Abstract

The present application relates to the technical field of data analysis, and particularly relates to an internet user portrait generation method based on big data, which comprises the following steps: obtaining the search times and search depth of a target data type obtained by browsing a webpage in a user monitoring time period, obtaining the search progression performance of the target data type, combining the interlaced search condition of other data types in the search process of the target data type, and obtaining the browsing concentration degree of the target data type; obtaining the increase degree of the browsing concentration degree based on the change trend of the browsing concentration degree of the target data type in multiple continuous monitoring time periods; and according to the increase degree of the browsing concentration degree, combining the browsing concentration degree of the target data type and the latest search time interval of the target data type, correctly mining the portrait importance degree of each target data type of the user, and improving the accuracy of the user portrait when performing the portrait of the user according to the portrait importance degree of each target data type.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data analysis, and particularly relates to an internet user portrait generation method based on big data. BACKGROUND

[0002] The internet user portrait based on big data can accurately identify user needs, preferences and behavior patterns, helping enterprises optimize product design, personalized recommendation and precision marketing, and significantly improving user experience and business conversion rate. At the same time, through data mining of potential risks (such as fraud and abnormal behavior), the platform security and compliance can be enhanced. In the competitive digital ecosystem, user portrait is a core tool for data-driven decision-making and cost reduction and efficiency improvement, and is also the basis for building intelligent services, having bidirectional value for enterprises and users.

[0003] The traditional method is: through the number of browsing times or the time length of different types of data of the user, the importance of the portrait of various types of data is determined, so as to realize the portrait of the user. However, when determining the importance of the portrait of various types of data, the influence of the frequent change of different types of data on the importance of the portrait of the user is not considered, so that the importance of the portrait of the user cannot be correctly mined, thereby affecting the accuracy of the portrait of the user. SUMMARY

[0004] In order to solve the technical problem that the existing user portrait method affects the accuracy of the portrait, the purpose of the present application is to provide an internet user portrait generation method based on big data, and the technical solution adopted is as follows:

[0005] The present application provides an internet user portrait generation method based on big data, comprising:

[0006] Obtaining the search times and search depth of the target data type obtained by browsing the webpage in the user monitoring time period to obtain the search progression performance of the target data type;

[0007] According to the search progression performance, combined with the interlaced search situation of other data types in the search process of the target data type, the browsing concentration degree of the target data type is obtained;

[0008] Based on the change trend of the browsing concentration degree of the target data type in a plurality of continuous monitoring time periods, the increase degree of the browsing concentration degree is obtained;

[0009] According to the increase degree of the browsing concentration degree, combined with the browsing concentration degree of the target data type and the recent search time interval of the target data type, the importance of the portrait of the target data type is obtained.

[0010] In an exemplary embodiment, the search progression performance acquisition process comprises:

[0011] obtaining a proportion of times of search times of the target data type;

[0012] obtaining a search depth difference between the search depth and a preset search depth;

[0013] obtaining a search progression performance of the target data type according to the proportion of times and the search depth difference, the search progression performance being proportional to the proportion of times and inversely proportional to the search depth difference.

[0014] In an exemplary embodiment, the obtaining process of the interleaved search case includes:

[0015] obtaining a data type number of other data types interleaved in search between any two search processes of the target data type and an interleaved search duration;

[0016] obtaining a concentration coefficient corresponding to the any two search processes according to the data type number and the interleaved search duration, the concentration coefficient being inversely proportional to the data type number and the interleaved search duration;

[0017] fusing concentration coefficients corresponding to all the any two search processes to obtain a comprehensive concentration coefficient.

[0018] In an exemplary embodiment, the obtaining process of the browsing concentration of the target data type includes:

[0019] fusing the comprehensive concentration coefficient and the search progression performance to obtain the browsing concentration.

[0020] In an exemplary embodiment, the obtaining process of the increase degree of the browsing concentration includes:

[0021] obtaining a proportion of monitoring time periods of a first monitoring time period, the first monitoring time period being a monitoring time period of the browsing concentration of the target data type greater than an adjacent previous monitoring time period;

[0022] obtaining an average value of differences of the browsing concentration of the target data type between all the first monitoring time periods and adjacent previous monitoring time periods thereof;

[0023] obtaining the increase degree of the browsing concentration according to the proportion of monitoring time periods and the average value of differences, the increase degree of the browsing concentration being proportional to the proportion of monitoring time periods and the average value of differences.

[0024] In an exemplary embodiment, the monitoring time period comprises a plurality of historical monitoring time periods and a current monitoring time period arranged in time sequence, and the latest search time interval of the target data type is a time interval between an end time of a search of the target data type in a target historical monitoring time period and an end time of the current monitoring time period; the target historical monitoring time period is a previous monitoring time period adjacent to the current monitoring time period.

[0025] The image importance is proportional to the increase degree of the browsing concentration and the browsing concentration of the target data type in the current monitoring time period, and is inversely proportional to the latest search time interval.

[0026] In an exemplary embodiment, the monitoring time period comprises a plurality of historical monitoring time periods and a current monitoring time period arranged in time sequence;

[0027] The target data type acquisition process comprises:

[0028] Based on the browsing concentration, the initial data types are divided into focused data types and non-focused data types; the initial data types comprise the target data types, and the target data types comprise the focused data types;

[0029] The browsing concentration difference of the focused data types before and after each non-focused data type is excluded from the current monitoring time period is obtained;

[0030] According to the interlaced search of each non-focused data type in the focused data type search process in the current monitoring time period, and in combination with the browsing concentration difference, the attention transfer degree of each non-focused data type to the focused data types is obtained;

[0031] According to the attention transfer degree and the browsing concentration of each non-focused data type, the image newly considered degree of each non-focused data type is obtained;

[0032] The non-focused data type corresponding to the image newly considered degree greater than a preset image newly considered degree threshold value is taken as the target data type.

[0033] In an exemplary embodiment, the attention transfer degree acquisition process comprises:

[0034] The search frequency proportion of each non-focused data type in the focused data type search process in the current monitoring time period is obtained;

[0035] The search time length proportion of each non-focused data type in the focused data type search process in the current monitoring time period is obtained;

[0036] According to the search frequency proportion, the search time length proportion and the browsing concentration change difference, a degree of attention transfer of each non-concentration data type to the concentration data type is obtained; the degree of attention transfer is directly proportional to the search frequency proportion, the search time length proportion and the browsing concentration change difference.

[0037] In one exemplary embodiment, the dividing of the initial data types into the concentration data type and the non-concentration data type based on the browsing concentration includes:

[0038] Comparing the browsing concentration of each initial data type in the current monitoring time period with a preset browsing concentration threshold value;

[0039] Determining the initial data type corresponding to the browsing concentration greater than the preset browsing concentration threshold value as the concentration data type.

[0040] In one exemplary embodiment, after obtaining the portrait importance degree of the target data type, the big data-based internet user portrait generation method further includes visualizing the user ID, the user data of the target data type and the portrait importance degree.

[0041] The present application has the following beneficial effects: the present application obtains the browsing concentration of the target data type by combining the search frequency and the search depth of the target data type obtained by the user in the monitoring time period and the interleaved search situation of other data types in the search process of the target data type, thereby obtaining the increase degree of the browsing concentration according to the dynamic change of the browsing concentration, and finally obtaining the portrait importance degree of the target data type according to the increase degree of the browsing concentration, the browsing concentration of the target data type and the latest search time interval of the target data type. The portrait importance degree of each target data type is closely related to the actual data situation of the target data type, thereby correctly mining the portrait importance degree of each target data type of the user. When performing the portrait of the user according to the portrait importance degree of each target data type, the accuracy of the user portrait can be improved. BRIEF DESCRIPTION OF DRAWINGS

[0042] Figure 1 is a flowchart of a big data-based internet user portrait generation method provided by one embodiment of the present application;

[0043] Figure 2 is a flowchart of obtaining of search progression performance provided by one embodiment of the present application;

[0044] Figure 3 is a flowchart of obtaining of interleaved search situation provided by one embodiment of the present application;

[0045] Figure 4is a flowchart of an acquisition process of a target data type provided by one embodiment of the present application;

[0046] Figure 5 is a flowchart of an acquisition process of a degree of attention diversion provided by one embodiment of the present application;

[0047] Figure 6 is a flowchart of an acquisition process of a degree of increase in browsing concentration provided by one embodiment of the present application. DETAILED DESCRIPTION

[0048] In order to further illustrate the technical means and effects taken by the present application to achieve the predetermined purposes, the specific embodiments, structures, features and effects of the present application are described in detail below in combination with the drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, the specific features, structures or characteristics in one or more embodiments can be combined in any suitable form.

[0049] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the present application belongs. The data information collected by the present application is obtained with full authorization.

[0050] The present embodiment provides an internet user portrait generation method based on big data. The applicable scenario is that the internet user portrait based on big data can accurately identify user demand, preference and behavior pattern, help enterprises optimize product design, personalized recommendation and precise marketing, and significantly improve user experience. At the same time, through data mining of potential risks (such as abnormal behavior), the platform safety and compliance can be enhanced.

[0051] As shown in Figure 1 The internet user portrait generation method based on big data provided by the present embodiment includes the following steps:

[0052] Step S1: acquiring the search times and search depth of the target data type obtained by browsing web pages in the user monitoring time period, to obtain the search progression performance of the target data type;

[0053] Step S2: according to the search progression performance, combining the interleaved search situation of other data types in the search process of the target data type, to obtain the browsing concentration of the target data type;

[0054] Step S3: based on the change trend of the browsing concentration of the target data type in multiple continuous monitoring time periods, to obtain the degree of increase in browsing concentration;

[0055] Step S4: According to the increase degree of browsing concentration, combining the browsing concentration of the target data type and the recent search time interval of the target data type, the portrait importance degree of the target data type is obtained.

[0056] The implementation process of each step is described in detail as follows in combination with the drawings.

[0057] Step S1: The search times and search depth of the target data type obtained by browsing web pages in the user monitoring time period are obtained, and the search progression performance of the target data type is obtained.

[0058] The embodiment processes the relevant information of the data type obtained by the user browsing the web page, and obtains the portrait importance degree of the corresponding data type, so as to realize the portrait operation on the user. Therefore, the user involved in the embodiment is any user. After obtaining the authorization of the user, the browsing history records (including the URL, access ip, access time, page title, etc. of the page) of the required date range are obtained through the BrowsingHistoryView cross-browser tool, and are exported in the form of a table. The Service Worker is set to monitor the page closing event to obtain the dwell time of the user on different pages, and the dwell time is stored in correspondence with the URL and other data of the browsed page.

[0059] The embodiment analyzes the browsing data in multiple monitoring time periods, wherein the multiple monitoring time periods include multiple historical monitoring time periods and a current monitoring time period arranged in time sequence. In the two adjacent monitoring time periods, the end time of the former monitoring time period and the start time of the latter monitoring time period are two adjacent times. The duration of each monitoring time period is equal, and the specific duration of each monitoring time period is set by actual needs, such as 1 week for each monitoring time period. The end time of the current monitoring time period is set as the current time. The following is described by taking any monitoring time period as an example.

[0060] The user browses the web page multiple times in the monitoring time period, and multiple data information can be obtained. According to the preset data type division mechanism, the data type of each obtained data information is obtained, and then multiple data types can be obtained. For example, if a specific brand of mobile phone is browsed, the corresponding data type division manner is: mobile phone specific model→mobile phone brand→mobile phone→electronic device, and the corresponding data type is electronic device; if a coat is browsed, the corresponding data type division manner is: coat→shirt→clothing, and the corresponding data type is clothing. Therefore, multiple different types of product information can be obtained, such as electronic devices, clothing, and different data types.

[0061] The various data types obtained by the user in the monitoring time period are defined as various initial data types. In an exemplary embodiment, the initial data types can be directly used as the target data types required by the present embodiment, and subsequent analysis is directly performed based on the data information of the initial data types, without the target data type screening process described below. As a more optimal implementation, the initial data types can also be screened to obtain target data types that are more closely related to the user portrait requirements. Then, the initial data types include the target data types, and the target data types are part of the initial data types.

[0062] In an exemplary embodiment, the target data types are obtained by screening the initial data types. Since the target data types are screened, the browsing concentration of each initial data type needs to be obtained. Therefore, the browsing concentration of each initial data type needs to be obtained first. It should be understood that since the target data types are part of the initial data types, the browsing concentration of each target data type is obtained at the same time as the browsing concentration of each initial data type is obtained.

[0063] For any initial data type, the user can obtain multiple data information of the initial data type by browsing the webpage in the monitoring time period, i.e., the initial data type corresponds to multiple browsing processes, and the data type of the data information obtained each time is the initial data type, such as the user browsing the shopping webpage of the electronic device multiple times in the monitoring time period, and obtaining a data information each time. Accordingly, the search times (i.e., the browsing times) of the initial data type in the monitoring time period are obtained, as well as the search depth of each search. The search times refer to the browsing times of the initial data type recorded from the browsing of the data information, and the search times of the initial data type are obtained. The search depth refers to the minimum classification of the data obtained each time, and the more specific the minimum classification is, the deeper the search depth is. For example, the search depth corresponding to the specific model of the mobile phone is deeper than the search depth corresponding to the brand of the mobile phone. Thus, based on the two information, the search progression of the initial data type is obtained. In an exemplary embodiment, as shown in FIG. 8, a specific acquisition process of the search progression is given as follows: Figure 2

[0064] Step S1-1: Obtain the proportion of the number of search times of the initial data type.

[0065] Since multiple initial data types are obtained in the monitoring time period, the search times of each initial data type can be obtained. Then, the sum of the search times of all the obtained initial data types is calculated, and then the ratio of the search times of the initial data type to the sum of the search times is calculated as the proportion of the number of search times of the initial data type. ​

[0066] Step S1-2: Obtain the search depth difference between each search depth and the preset search depth.

[0067] In order to facilitate calculation, the search depth level corresponding to the initial data type is quantized. The more specific the level corresponding to the search depth is, that is, the smaller it is, the greater the value of the search depth is. The value of the search depth is the value of the corresponding level. For example, for the four levels of mobile phone specific model, mobile phone brand, mobile phone, and electronic device, the search depth of electronic device is 1, the search depth of mobile phone is 2, the search depth of mobile phone brand is 3, and the search depth of mobile phone specific model is 4.

[0068] By using the above process, the search depth obtained by each search of the initial data type is quantized to obtain the quantized value of the search depth obtained by each search.

[0069] And for the above quantization, for the initial data type, a preset search depth is set, which is the minimum level allowed for the initial data type. For example, for mobile phone specific model, the quantized value of the preset search depth is 4.

[0070] Then, the search depth difference between each search depth and the preset search depth, that is, the absolute value of the difference between the quantized value of each search depth and the quantized value of the preset search depth, is obtained. Then, the average value of the absolute value of the difference between the quantized value of all search depths of the initial data type in the monitoring time period and the quantized value of the preset search depth is calculated as the search depth difference of the initial data type in the monitoring time period. The search depth difference of the initial data type represents the gap between the preset maximum search depth level. The smaller the search depth difference is, the smaller the gap between the preset maximum search depth level is, and the deeper the search depth is.

[0071] Step S1-3: Obtain the search progression performance of the initial data type according to the proportion of times and the search depth difference.

[0072] According to the proportion of times of the search times of the initial data type and the search depth difference of the initial data type, the search progression performance of the initial data type is obtained.

[0073] When the proportion of times of the search times of the initial data type is greater, the search depth difference of the initial data type is smaller, the search progression performance of the initial data type is more obvious, the preference of the user to the data of the initial data type may be greater, and the user portrait should be retained. Then, the search progression performance is proportional to the proportion of times and inversely proportional to the search depth difference.

[0074] In an exemplary embodiment, a quantization method of search progression performance is given as follows:

[0075] ;

[0076] wherein, represents the search progressive performance of the jthinitial data type, represents the search depth difference of the jthinitial data type, exp represents the exponential function with the natural constant e as the base, represents the negative correlation normalization of In the embodiment, the negative correlation normalization is in this way. represents the proportion of the number of searches of the jthinitial data type. The calculation formula is essentially a weighted sum of two parameters affecting the search progressive performance, and the weight is 0.5.

[0077] Step S2: According to the search progressive performance, combined with the interleaved search situation of other data types in the search process of the target data type, the browsing concentration of the target data type is obtained.

[0078] For the initial data type, the search process of the user on the initial data type may be interleaved with the search of other data types, that is, the user also browses some web pages of other data types in the search process of the initial data type. Then, according to the search progressive performance of the initial data type, combined with the interleaved search situation of other data types in the search process of the initial data type, the browsing concentration of the initial data type is obtained.

[0079] In an exemplary embodiment, as shown in Figure 3 A specific acquisition process of the interleaved search situation is as follows:

[0080] Step S2-1: Obtain the data type quantity of other data types interleaved in the search process and the interleaved search duration.

[0081] Since the user will search the initial data type for multiple times in the monitoring time period, for any two adjacent search processes of the initial data type, the data type quantity of other data types interleaved in the time period between the two adjacent search processes of the initial data type and the interleaved search duration are obtained. Among them, the various data types searched by the user in the time period between the two adjacent search processes of the initial data type are obtained, so as to obtain the data type quantity of the various data types searched by the user. The search duration of each data type searched by the user in the time period between the two adjacent search processes of the initial data type is obtained, and then added to obtain the sum of the search durations, which is the interleaved search duration.

[0082] Step S2-2: According to the data type quantity and the interlaced search duration, obtain the concentration coefficient corresponding to any two search processes.

[0083] When the data type quantity of other data types interlaced in the search is smaller, it indicates that the user's interest in other data types is lower in the time period between any two search processes adjacent to the initial data type, and the concentration coefficient corresponding to the two adjacent search processes for the initial data type is higher. When the interlaced search duration of other data types interlaced in the search is shorter, it indicates that the user's interest in other data types is lower in the time period between any two search processes adjacent to the initial data type, and the concentration coefficient corresponding to the two adjacent search processes for the initial data type is higher. The concentration coefficient is inversely proportional to the type quantity and the interlaced search duration.

[0084] In an exemplary embodiment, the product of the type quantity and the interlaced search duration is calculated, and negative correlation normalization is performed, and the result is the concentration coefficient corresponding to the two search processes for the initial data type.

[0085] Step S2-3: Fuse the concentration coefficients corresponding to all any two search processes to obtain the comprehensive concentration coefficient of the initial data type.

[0086] Through step S2-2, the concentration coefficient corresponding to any two adjacent search processes for the initial data type is obtained, and the concentration coefficients corresponding to all any two search processes of the initial data type are fused, specifically, the average value of the concentration coefficients is calculated, and the result is the comprehensive concentration coefficient of the initial data type.

[0087] The larger the comprehensive concentration coefficient is, the more reference value the initial data type has in user profiling, that is, the greater the browsing concentration of the initial data type is. The greater the search progression performance of the user is, the more reference value the initial data type has in user profiling, that is, the greater the browsing concentration of the initial data type is.

[0088] Fuse the comprehensive concentration coefficient of the initial data type and the search progression performance to obtain the browsing concentration of the initial data type. The browsing concentration of the initial data type is proportional to the comprehensive concentration coefficient of the initial data type and the search progression performance. In an exemplary embodiment, a specific fusion method is given as follows: the comprehensive concentration coefficient of the initial data type and the search progression performance of the initial data type are weighted and summed, and the weight is 0.5, that is, the average value of the comprehensive concentration coefficient of the initial data type and the search progression performance of the initial data type is calculated, and the result is the browsing concentration of the initial data type.

[0089] With the above process, the browsing concentration of each initial data type is obtained. Then, the target data type can be filtered from each initial data type based on the browsing concentration of each initial data type. In order to improve the reliability of the target data type filtering, the target data type is obtained from the current monitoring time period.

[0090] In an exemplary embodiment, as shown in Figure 4 A specific obtaining process of the target data type is as follows:

[0091] Step S2-4: Based on the browsing concentration, the initial data types are divided into focused data types and non-focused data types.

[0092] With the above method, the browsing concentration of each initial data type in the current monitoring time period is obtained. In order to improve the reliability of the subsequent screening, after obtaining the browsing concentration of each initial data type in the current monitoring time period, the maximum value and the minimum value of the browsing concentration are obtained, and then the maximum value and the minimum value normalization method is used to normalize the browsing concentration of each initial data type in the current monitoring time period. The browsing concentration of each initial data type in the current monitoring time period used in the subsequent comparison is the result of the normalization by the maximum value and the minimum value normalization method.

[0093] A browsing concentration threshold is preset, which is used to determine whether the browsing concentration of each initial data type in the current monitoring time period is high, so as to realize the screening of focused data types and non-focused data types. The numerical range of the preset browsing concentration threshold is 0-1, and the specific value of the preset browsing concentration threshold is set by actual needs under the premise of meeting the comparison. In this embodiment, 0.7 is taken as an example.

[0094] The browsing concentration of each initial data type in the current monitoring time period is compared with the preset browsing concentration threshold, and the initial data type corresponding to the browsing concentration greater than the preset browsing concentration threshold is determined as the focused data type. Correspondingly, the other initial data types except the focused data types are non-focused data types, so that the initial data types are divided into focused data types and non-focused data types. The focused data types are taken as the target data types, then the target data types include the focused data types, and the purpose of the subsequent steps is to filter a part of the data types from the non-focused data types as the target data types, so as to realize the addition and update of the target data types.

[0095] Step S2-5: Obtain the browsing concentration change difference of the focused data types before and after each non-focused data type is excluded from the current monitoring time period.

[0096] The following is an example of any one non-focused data type. The browsing focus of the non-focused data type is obtained. Then the non-focused data type is excluded from the current monitoring time period, so that the distribution of data types in the current monitoring time period is changed. Since the browsing focus is related to its own data type and also related to other data types in the current monitoring time period, the browsing focus of each data type will change after the non-focused data type is excluded from the current monitoring time period.

[0097] For any one focused data type, the browsing focus of the focused data type before the non-focused data type is excluded from the current monitoring time period is obtained. The browsing focus of the focused data type after the non-focused data type is excluded from the current monitoring time period is obtained. Then the difference between the browsing focus of the focused data type before and after the non-focused data type is excluded from the current monitoring time period is obtained. The difference between the browsing focus of the focused data type before and after the non-focused data type is excluded from the current monitoring time period is specifically the absolute value of the difference between the browsing focus of the focused data type before and after the non-focused data type is excluded from the current monitoring time period.

[0098] Each focused data type is calculated in the above manner, so that for the non-focused data type, the browsing focus change difference corresponding to each focused data type is obtained. Then for all focused data types, the average of the browsing focus change difference is calculated to obtain the browsing focus change difference corresponding to the non-focused data type, and the calculation formula is as follows:

[0099] ;

[0100] Wherein, represents the browsing focus change difference corresponding to the gth non-focused data type, represents the browsing focus of the ith focused data type before the gth non-focused data type is excluded from the current monitoring time period, represents the browsing focus of the ith focused data type after the gth non-focused data type is excluded from the current monitoring time period, represents the number of focused data types.

[0101] The larger the browsing focus change difference, the more serious the influence of the non-focused data type on the overall browsing focus of all data types before and after the non-focused data type is excluded, the more important the non-focused data type, and the higher the degree of attention transfer of the non-focused data type.

[0102] Step S2-6: According to the search situation of each non-focused data type in the current monitoring time period, combined with the difference in browsing concentration, the degree of attention transfer of each non-focused data type to the focused data type is obtained.

[0103] In the current monitoring time period, it is possible to search for non-focused data types during the search for focused data types. The more serious the search situation of non-focused data types, the more serious the attention transfer of non-focused data types to focused data types. Combined with the difference in browsing concentration, the degree of attention transfer of non-focused data types to focused data types is obtained.

[0104] In one exemplary embodiment, as shown in Figure 5 A specific process for obtaining the degree of attention transfer is as follows:

[0105] Step S2-6-1: Obtain the search frequency ratio of each non-focused data type in the current monitoring time period during the search for focused data types.

[0106] The gth non-focused data type represents any non-focused data type, and the ith focused data type represents any focused data type. In the current monitoring time period, the search frequency of the gth non-focused data type during the search for the ith focused data type is obtained, so that the search frequency of the gth non-focused data type during the search for each focused data type is obtained. The sum of the search frequencies of the gth non-focused data type for all focused data types is calculated. Then the total sum of the search frequencies of all non-focused data types for all focused data types is calculated, and finally the ratio of the sum of the search frequencies of the gth non-focused data type to the total sum of the search frequencies is calculated as the search frequency ratio of the gth non-focused data type.

[0107] Therefore, the greater the search frequency ratio of the gth non-focused data type, the higher the search degree of the gth non-focused data type during the search for focused data types, and the higher the degree of attention transfer of the gth non-focused data type. The search frequency ratio is proportional to the degree of attention transfer.

[0108] Step S2-6-2: Obtain the search time ratio of each non-focused data type in the current monitoring time period during the search for focused data types.

[0109] In the current monitoring time period, in the search process of the i-th focused data type, the search duration of the g-th non-focused data type is obtained, so as to obtain the search duration of the g-th non-focused data type in each focused data type search process. The sum value of the search duration of the g-th non-focused data type is calculated for all focused data types. Then the sum value of the search duration of all non-focused data types for all focused data types is calculated, and finally the ratio of the sum value of the search duration of the g-th non-focused data type to the sum value of the search duration is calculated as the search duration ratio of the g-th non-focused data type.

[0110] Therefore, the greater the search duration ratio of the g-th non-focused data type, the higher the degree of interleaved search of the g-th non-focused data type in the focused data type search process, and the higher the attention transfer degree of the g-th non-focused data type. The search duration ratio is proportional to the attention transfer degree.

[0111] Step S2-6-3: According to the search frequency ratio, the search duration ratio and the browsing focus degree change difference, the attention transfer degree of each non-focused data type to the focused data type is obtained.

[0112] According to the search frequency ratio, the search duration ratio and the browsing focus degree change difference of the g-th non-focused data type, the attention transfer degree of the g-th non-focused data type to the focused data type is obtained. The attention transfer degree is proportional to the search frequency ratio, the search duration ratio and the browsing focus degree change difference.

[0113] In an exemplary embodiment, a specific quantification method of the attention transfer degree of the g-th non-focused data type is given as follows: the product of the search frequency ratio, the search duration ratio and the browsing focus degree change difference of the g-th non-focused data type is calculated, and the result is the attention transfer degree of the g-th non-focused data type to the focused data type.

[0114] Step S2-7: According to the attention transfer degree and the browsing focus degree of each non-focused data type, the portrait addition consideration degree of each non-focused data type is obtained.

[0115] The higher the attention transfer degree of the non-focused data type, the more important it is to consider adding its portrait, and the higher the portrait addition consideration degree; the higher the browsing focus degree of the non-focused data type, the more important it is to consider adding its portrait, and the higher the portrait addition consideration degree. Therefore, the portrait addition consideration degree is proportional to the attention transfer degree and the browsing focus degree.

[0116] In an exemplary embodiment, a specific quantification manner of the portrait newly-added consideration degree of the gth non-focused data type is given as follows: the product of the attention shift degree of the gth non-focused data type and the browsing concentration of the gth non-focused data type is calculated, and then the product is normalized, and the normalized result is the portrait newly-added consideration degree of the gth non-focused data type. The normalization manner here is: the maximum value and the minimum value in the products of the attention shift degrees and the browsing concentrations of various non-focused data types are obtained, and then the product of the attention shift degree of the gth non-focused data type and the browsing concentration of the gth non-focused data type is normalized by using the maximum value and the minimum value normalization manner.

[0117] Step S2-8: taking the non-focused data type corresponding to the portrait newly-added consideration degree greater than the preset portrait newly-added consideration degree threshold value as the target data type.

[0118] The embodiment presets a portrait newly-added consideration degree threshold value, which is used to determine whether the portrait newly-added consideration degree of various non-focused data types is high, so as to complete the screening of non-focused data types. It should be understood that the numerical range of the preset portrait newly-added consideration degree threshold value is 0-1, and the specific value of the preset portrait newly-added consideration degree threshold value is flexibly set by the implementer under the premise of meeting the above-mentioned determination needs, such as 0.7.

[0119] The sizes of the portrait newly-added consideration degrees of various non-focused data types and the preset portrait newly-added consideration degree threshold value are compared, the non-focused data type corresponding to the portrait newly-added consideration degree greater than the preset portrait newly-added consideration degree threshold value is obtained, and the non-focused data type obtained here is taken as the target data type, so as to complete the addition and update of the target data type.

[0120] Therefore, in addition to directly screening through the browsing concentration above, for the data types not meeting the condition of the browsing concentration, the above-mentioned process is used to complete the addition of the target data type, so as to improve the accuracy of obtaining the target data type and avoid causing the omission of the data type.

[0121] Step S3: obtaining the browsing concentration increase degree based on the change trend of the browsing concentration of the target data type in a plurality of continuous monitoring time periods.

[0122] Since the attention of the user to different things is usually constantly changing, the user's portrait should also be constantly changed accordingly. When the user's browsing concentration on a data type increases in multiple consecutive monitoring time periods, the user's preference for this data type should be increased, i.e., the portrait importance should be increased, and vice versa. And the closer the last moment when this data type is searched by the user to the present, and the greater the corresponding browsing concentration, the more the user's attention to this data type should be emphasized.

[0123] The browsing concentration of each target data type in each monitoring time period is obtained by step S2. The following is described by taking any one target data type as an example. Then, the browsing concentration of the target data type in multiple consecutive monitoring time periods is obtained, and then the browsing concentration of the target data type in each monitoring time period is arranged in time sequence, so as to obtain the browsing concentration sequence of the target data type. According to the change trend of the browsing concentration in the browsing concentration sequence of the target data type, the increase degree of the browsing concentration of the target data type is obtained.

[0124] In an exemplary embodiment, as shown in Figure 6 A specific process for obtaining the increase degree of the browsing concentration is given as follows:

[0125] Step S3-1: Obtain the monitoring time period number proportion of the first monitoring time period.

[0126] This embodiment sets a judgment logic: for any monitoring time period, if the browsing concentration of the target data type of the monitoring time period is greater than the browsing concentration of the target data type of the adjacent previous monitoring time period, it means that the browsing concentration of the target data type of the monitoring time period is in an increasing state, and then the monitoring time period is defined as the first monitoring time period. That is, the first monitoring time period refers to the monitoring time period whose browsing concentration of the target data type is greater than that of the adjacent previous monitoring time period.

[0127] All monitoring time periods are traversed, and each monitoring time period is screened to determine whether it is the first monitoring time period. Then the number of the first monitoring time period is obtained. The ratio of the number of the first monitoring time period to the total number of the monitoring time periods is calculated as the monitoring time period number proportion of the first monitoring time period. The greater the monitoring time period number proportion of the first monitoring time period, the greater the increase degree of the browsing concentration of the target data type. The increase degree of the browsing concentration is proportional to the monitoring time period number proportion.

[0128] It should be understood that since there is no monitoring time period before the first monitoring time period in time sequence, the first monitoring time period does not participate in the comparison of the browsing concentration of the target data type with the adjacent previous monitoring time period.

[0129] Step S3-2: Obtain the average value of the difference of the browsing concentration of the target data type between each first monitoring time period and its adjacent previous monitoring time period.

[0130] For any one first monitoring time period, the difference of the browsing concentration of the target data type between the first monitoring time period and its adjacent previous monitoring time period is calculated. Then the average value of the difference of the browsing concentration of the target data type between each first monitoring time period and its adjacent previous monitoring time period is calculated. The greater the average value of the difference is, the more obvious the increase of the browsing concentration of the target data type is, i.e. the greater the increase degree of the browsing concentration of the target data type is. The increase degree of the browsing concentration is proportional to the average value of the difference.

[0131] Step S3-3: Obtain the increase degree of the browsing concentration according to the proportion of the number of monitoring time periods and the average value of the difference.

[0132] According to the proportion of the number of monitoring time periods of the target data type and the average value of the difference corresponding to the target data type, the increase degree of the browsing concentration of the target data type is obtained. In an exemplary embodiment, a quantitative method of the increase degree of the browsing concentration is given as follows: the product of the proportion of the number of monitoring time periods of the target data type and the average value of the difference corresponding to the target data type is calculated, and then normalized to obtain the result as the increase degree of the browsing concentration of the target data type. The normalization method here can be a sigmoid function. The greater the increase degree of the browsing concentration of the target data type is, the higher the portrait importance degree of the target data type is. The portrait importance degree is proportional to the increase degree of the browsing concentration.

[0133] Step S4: Obtain the portrait importance degree of the target data type according to the increase degree of the browsing concentration, combined with the browsing concentration of the target data type and the recent search time interval of the target data type.

[0134] In this embodiment, the target historical monitoring time period is set as the previous monitoring time period adjacent to the current monitoring time period. The search end time of the target data type in the target historical monitoring time period is obtained, so as to obtain the time interval between the search end time of the target data type in the target historical monitoring time period and the cutoff time of the current monitoring time period, which is the recent search time interval of the target data type. The shorter the recent search time interval of the target data type is, the closer the recent search of the target data type is to the current time, and the higher the portrait importance degree of the target data type is. Therefore, the portrait importance degree of the target data type is inversely proportional to the recent search time interval of the target data type.

[0135] Further, the higher the browsing concentration of the target data type in the current monitoring time period, the higher the portrait importance of the target data type. The portrait importance of the target data type is proportional to the browsing concentration of the target data type in the current monitoring time period.

[0136] In an exemplary embodiment, a quantitative manner of the portrait importance of the target data type is given as follows: the recent search time interval of the target data type is negatively correlated and normalized, then the product of the increase of the browsing concentration of the target data type, the browsing concentration of the target data type in the current monitoring time period and the negatively correlated and normalized recent search time interval of the target data type is calculated, and the obtained product is taken as the portrait importance feature of the target data type. Finally, the portrait importance feature of the target data type is normalized, and the normalized result is the portrait importance of the target data type. The normalization manner herein can be: the maximum value and the minimum value in the portrait importance features of the target data types are obtained, and then the maximum value and the minimum value normalization manner is adopted to normalize the portrait importance features of the target data types.

[0137] By using the above process, the portrait importance of each target data type is obtained. Then the user ID, the user data of each target data type and the portrait importance of each target data type are corresponded and stored in the database. The user data of each target data type can be called from the database by using the SQL query statement. Then the user ID, the user data of each target data type and the portrait importance of each target data type are visualized and displayed, such as shown in Table 1, which is an example of the user portrait visualization result.

[0138] Table 1

[0139]

[0140] In addition, the target data type with the portrait importance greater than a preset value, such as 0.5, can also be selected as the reference of the user's favorite in the user portrait. Meanwhile, the KMP (Knuth-Morris-Pratt, string matching algorithm) matching algorithm is used to match the name, ID, gender and other data from the user's personal information field, so as to correspond and store the favorite data type of the user with the user information.

[0141] It should be noted that the above-mentioned sequence of the embodiments of the present application is only for description, and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multi-task processing and parallel processing are also possible or can be advantageous.

[0142] The various embodiments described in this specification are presented by way of example, and each embodiment is not inherently more important than any other embodiment.

Claims

1. A method for generating internet user profiles based on big data, characterized in that, include: The number of searches and search depth of the target data type obtained from browsing web pages during the user monitoring period are obtained to obtain the progressive performance of the search of the target data type. Based on the search progression behavior and the interleaving search of other data types during the search process for the target data type, the browsing focus of the target data type is obtained. Based on the changing trend of browsing focus on target data types over multiple consecutive monitoring periods, the degree of increase in browsing focus is obtained. Based on the increase in browsing focus, combined with the browsing focus of the target data type and the most recent search time interval of the target data type, the importance of the target data type profile is obtained; The search depth is the smallest category of data obtained during each browsing session. The more specific the smallest category, the deeper the search depth. The process of obtaining the interleaved search situation includes: Get the number of other data types searched between any two adjacent search processes on the target data type, as well as the duration of the interleaved search; Based on the number of data types and the duration of interleaved searches, the focus coefficient corresponding to any two search processes is obtained; the focus coefficient is inversely proportional to both the number of types and the duration of interleaved searches. The overall focus coefficient is obtained by combining the focus coefficients corresponding to any two search processes.

2. The method for generating internet user profiles based on big data as described in claim 1, characterized in that, The process of obtaining the search progression representation includes: Obtain the percentage of searches for the target data type; Obtain the difference between the stated search depth and the preset search depth; Based on the percentage of times and the difference in search depth, the progressive search performance of the target data type is obtained. The progressive search performance is directly proportional to the percentage of times and inversely proportional to the difference in search depth.

3. The method for generating internet user profiles based on big data as described in claim 1, characterized in that, The process of obtaining browsing focus for the target data type includes: The browsing focus score is obtained by combining the overall focus score and the search progression performance.

4. The method for generating internet user profiles based on big data as described in claim 1, characterized in that, The process of obtaining the degree of increase in browsing focus includes: Obtain the percentage of monitoring time periods in the first monitoring time period, where the first monitoring time period refers to the monitoring time period in which the browsing focus of the target data type is greater than that of the adjacent previous monitoring time period. Obtain the average of the differences in browsing focus for all target data types in the first monitoring time period and the preceding adjacent monitoring time period; The degree of increase in browsing focus is obtained based on the average of the proportion of the monitored time periods and the difference. The degree of increase in browsing focus is directly proportional to both the proportion of the monitored time periods and the average of the difference.

5. The method for generating internet user profiles based on big data as described in claim 1, characterized in that, The monitoring time period includes multiple historical monitoring time periods arranged in chronological order and the current monitoring time period. The most recent search time interval of the target data type is the time interval between the end time of the search for the target data type in the target historical monitoring time period and the end time of the current monitoring time period. The target historical monitoring time period is the previous monitoring time period adjacent to the current monitoring time period. The importance of the profile is directly proportional to the increase in browsing focus and the browsing focus on the target data type during the current monitoring period, and inversely proportional to the most recent search time interval.

6. The method for generating internet user profiles based on big data as described in claim 1, characterized in that, The monitoring time period includes multiple historical monitoring time periods arranged in chronological order and the current monitoring time period; The process of obtaining the target data type includes: Based on browsing focus, the initial data types are divided into focused data types and non-focused data types; The initial data type includes the target data type, and the target data type includes the focused data type; Obtain the difference in browsing focus on focused data types before and after removing each non-focused data type from the current monitoring time period; Based on the interspersed search of each non-focused data type during the focused data type search process in the current monitoring period, and combined with the differences in browsing focus, the degree of attention shift from each non-focused data type to the focused data type is obtained. Based on the degree of attention shift and browsing focus for each non-focused data type, the degree of additional consideration for each non-focused data type profile is obtained; The non-focused data type corresponding to the new image consideration level that is greater than the preset new image consideration level threshold is taken as the target data type.

7. The method for generating internet user profiles based on big data as described in claim 6, characterized in that, The process of obtaining the degree of attention shift includes: Obtain the percentage of searches for each non-focused data type during the focused data type search process within the current monitoring period; Obtain the percentage of search time for each non-focused data type during the focused data type search process within the current monitoring time period; Based on the percentage of search counts, the percentage of search duration, and the difference in browsing focus, the degree of attention shift from the non-focused data type to the focused data type is obtained; the degree of attention shift is directly proportional to the percentage of search counts, the percentage of search duration, and the difference in browsing focus.

8. The method for generating internet user profiles based on big data as described in claim 6, characterized in that, Based on browsing focus, the initial data types are divided into focused data types and non-focused data types, including: Compare the browsing focus level of each initial data type during the current monitoring period with the preset browsing focus level threshold; The initial data type corresponding to the browsing focus level that is greater than the preset browsing focus level threshold is determined as the focus data type.

9. The method for generating internet user profiles based on big data as described in claim 1, characterized in that, After obtaining the profile importance of the target data type, the big data-based Internet user profile generation method further includes: visualizing the user ID, user data of the target data type, and profile importance.

Images