Visualization of user operation data and human-computer interaction recommendation method

Abstract

The present invention discloses a visualization and human-computer interaction recommendation method for user operation data, which belongs to the field of recommendation technology. The method comprises: determining each recommendation point in real time, and performing association classification according to each recommendation point to form each association tree; generating a recommendation principle diagram according to each association tree; identifying the user's operation data in real time, performing statistical analysis on the operation data according to the recommendation principle diagram, determining the recommendation weight value of each recommendation point in the recommendation principle diagram, and supplementing the recommendation weight value of each recommendation point in the recommendation principle diagram; generating a corresponding recommendation detail table in real time according to the recommendation principle diagram; marking the current recommendation principle diagram as a data display diagram, and displaying the data display diagram to the user in real time; making real-time recommendations to the user according to the data display diagram, and the user interactively adjusting the data display diagram according to the recommendation result; reasonably organizing and displaying a large number of recommendation points through level-by-level classification; reducing the difficulty for the user to understand the recommendation principle.

Description

Technical Field

[0001] The present invention belongs to the technical field of recommendation, and in particular relates to a visualization and human-computer interaction recommendation method for user operation data. Background Art

[0002] In current recommendation systems, although the technology is changing with each passing day, providing users with more personalized and accurate content recommendations, many users are still confused about the inner workings of recommendation systems. This opacity mainly stems from the complexity of recommendation algorithms and the lack of intuitive user interfaces to show how these algorithms work. Therefore, when users face unsatisfactory recommendation results, they often feel at a loss as to where to start and find it difficult to adjust the recommendation parameters to optimize the recommendation results.

[0003] A key problem with existing recommendation systems is that they usually rely on multiple recommendation algorithms and strategies that predict user interests and preferences through complex computational models. These models may include collaborative filtering, content recommendation, association rule-based recommendation, etc. Each method has its own unique advantages and limitations. However, for most users, these technical details are obscure and difficult to understand. They are more concerned about how to quickly adjust the recommendation results according to their needs and avoid continuous recommendation of uninteresting content.

[0004] Based on this, in order to solve the recommendation problem, the present invention provides a visualization and human-computer interaction recommendation method for user operation data. Summary of the invention

[0005] In order to solve the problems existing in the above solutions, the present invention provides a visualization and human-computer interaction recommendation method for user operation data.

[0006] The purpose of the present invention can be achieved through the following technical solutions:

[0007] The visualization and human-computer interaction recommendation method of user operation data includes:

[0008] Step 1: Determine each recommendation point in real time, perform association classification according to each recommendation point, and form each association tree; generate a recommendation principle diagram according to each association tree.

[0009] Furthermore, the method for performing association analysis based on each recommended point is:

[0010] Establishing a correlation priority model, and evaluating the correlation relationship between the recommendation points through the correlation priority model;

[0011] Classification is performed according to the association relationship between the recommendation points to form a plurality of tree-like material classes, the association relationship between the recommendation points in each tree-like material class is identified, and an association tree corresponding to each tree-like material class is formed according to each association relationship.

[0012] Furthermore, the expression of the association priority model is:

[0013] ;

[0014] Where: (q, p) is the input data, q and p represent the two recommended points for comparative analysis; q→p means that the association level of q is higher than that of p; q←p means that the association level of q is lower than that of p; the output data is the association priority value GL(q, p), and the association priority values ​​are 0, 1, and -1;

[0015] When the association priority value is 0, it means that the two recommendation points have no association;

[0016] When the association priority value is 1, it means that the two recommended points are related, and the association level of q is higher than that of p;

[0017] When the association priority value is -1, it means that the two recommended points are associated, and the association level of q is lower than that of p.

[0018] Furthermore, the method for generating a recommended schematic diagram according to each association tree is:

[0019] Identify each association level corresponding to each association tree, generate each initial layer according to each association level, and the association level of each recommended point in the initial layer is the same; identify the number of recommended points corresponding to each initial layer, and mark them as layer values;

[0020] The initial layers are adjusted according to preset display limits and layer values ​​to obtain display layers, and the display layers are visualized according to preset display methods to obtain a recommended schematic diagram.

[0021] Furthermore, the method for adjusting each initial layer according to the display limit value and each layer value includes:

[0022] Determine a reference layer in each of the initial layers, identify a layer value corresponding to the reference layer, and mark it as a reference value;

[0023] When the base value is not greater than the display limit, the base layer is not processed;

[0024] When the benchmark value is greater than the display limit, the benchmark layer is merged;

[0025] Sort and distribute the baseline layer and the remaining initial layers according to the correlation levels of the corresponding recommended points;

[0026] Compare the layer values ​​of the remaining initial layers with the display limit value, determine the initial layers that need to be merged, and merge the initial layers that need to be merged;

[0027] Mark each initial layer and base layer as a display layer.

[0028] Furthermore, the method for merging the reference layers includes:

[0029] Calculate the difference between the benchmark value and the display limit value, and mark it as the display difference; perform simulated upgrade according to the display difference, and obtain various simulated upgrade methods;

[0030] Acquire the user's operation records in real time, perform matching analysis on each of the simulated upgrade methods according to the operation records, and determine the target upgrade method;

[0031] The base layers are merged according to the target upgrading method.

[0032] Furthermore, the method for matching and analyzing each simulated upgrade mode according to the operation record is as follows:

[0033] Identify each recommended point in the reference layer and the record weight of each recommended point according to the operation record;

[0034] Process the reference layer according to each of the simulation upgrading methods, and mark the processed reference layer as a simulation layer; mark the record weight of each recommended point in the simulation layer as ηi, where i=1, 2, ..., n, and n is the number of recommended points in the simulation layer;

[0035] The simulation value of each simulation layer is calculated according to the simulation value formula; the simulation value formula is:

[0036] ;

[0037] Where: PA is the analog value;

[0038] The simulation upgrading method corresponding to the simulation layer with the largest simulation value is selected as the target upgrading method.

[0039] Step 2: Identify the user's operation data in real time, perform statistical analysis on the operation data according to the recommendation schematic diagram, determine the recommendation weight value of each recommendation point in the recommendation schematic diagram, and supplement the recommendation weight value of each recommendation point in the recommendation schematic diagram; generate a corresponding recommendation detail table in real time according to the recommendation schematic diagram; mark the current recommendation schematic diagram as a data display diagram, and display the data display diagram to the user in real time.

[0040] Furthermore, the method of statistically analyzing the operation data according to the recommended schematic diagram includes:

[0041] Determine the operation record of each recommended point according to the operation data and the recommended principle diagram, preset a unit period, and generate a record curve of each recommended point according to the unit period and each operation record, wherein the horizontal axis of the record curve is the time representative axis corresponding to each unit period; the vertical axis is the number of operations counted in the corresponding unit period, marked as a unit value;

[0042] Set a time correction function; correct the recorded curve of each recommended point according to the time correction function, mark the corrected recorded curve as a unit curve, fit the unit curve, and obtain the unit function;

[0043] The unit function of each recommended point in the same display layer is marked as DH(T)j, where T is the horizontal axis variable of the unit function, j=1, 2, ..., m, and m is the number of recommended points in the display layer;

[0044] Calculate the recommended weight value of each recommended point according to the recommended weight formula; the recommended weight formula is:

[0045] ;

[0046] Where: PKj is the recommended weight value;

[0047] .

[0048] Furthermore, when the data display graph does not meet the user's needs, the user sets the stage time and adjusts the correction coefficient corresponding to the time correction function outside the stage time to 0; the adjusted time correction function is used to perform correction to form a new data display graph.

[0049] Step three: making real-time recommendations to the user based on the data display graph, and the user interactively adjusts the data display graph based on the recommendation results.

[0050] Furthermore, when the user sets the recommendation weight value of the recommendation point in the data display chart, there are two additional function options. One additional function option is dynamic update and fixed duration, that is, no update is performed within the fixed time period, and after the fixed time period, it is updated as the operation data changes; the other additional function is fixed, that is, there is no subsequent change unless the user makes adjustments or uses another additional function option.

[0051] Compared with the prior art, the present invention has the following beneficial effects:

[0052] The present invention rationally organizes and displays numerous recommendation points by means of level-by-level classification. This design not only reduces the difficulty for users to understand the recommendation principle, but also enables users to quickly locate the recommendation points of interest. At the same time, users can flexibly adjust the displayed classification level through operations such as zooming in and out, further enhancing user participation and experience. In addition, users can adjust the weight of each recommendation point according to actual needs, thereby quickly correcting the recommendation trend and making the recommendation results more in line with the user's personalized needs. It not only solves the problems of users' difficulty in understanding the recommendation principle and adjusting the recommendation results in the existing recommendation system, but also provides new ideas and methods for the innovation and development of recommendation technology. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0054] Figure 1 The figure is a flow chart of the method of the present invention. DETAILED DESCRIPTION

[0055] The technical solution of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0056] like Figure 1 As shown, the visualization of user operation data and the human-computer interaction recommendation method include:

[0057] Step 1: The platform determines each recommendation point based on the available recommendable content. Recommendation points such as recommendation tags, keywords, categories, etc. are determined based on the application recommendation algorithm; associations are graded based on each recommendation point to form association trees; and recommendation principle diagrams are generated based on each association tree and visualization technology.

[0058] In one embodiment, recommendation points can be quickly identified based on existing data recognition and other technologies to determine the recommendation points in the platform; and newly added recommendation points can be updated in real time according to subsequent developments.

[0059] In one embodiment, the method for performing association analysis based on each recommendation point is: according to the existing technology management analysis, for example, there are three recommendation points of food, rice, and black rice, and according to their association relationship, an association tree corresponding to food-rice-black rice can be formed, corresponding to three association levels, food is the first level, rice is the second level, black rice is the third level, and so on. If there are corresponding recommendation points later, the corresponding levels are determined in turn. If there are fruit and apple recommendation points, the fruit is assigned to the second level and the apple is assigned to the third level according to the association relationship.

[0060] In one embodiment, the method for performing association analysis based on each recommendation point is:

[0061] Establish an association priority model, which is used to evaluate whether there is an association and a comparative hierarchical relationship between two recommendation points. For example, food and rice, food and fruit are associated, and food can contain rice and fruit, so the level of food is higher than that of rice and fruit, while there is no inclusion relationship between rice and fruit, and they are considered to be unrelated, but they are both below the level of food and are considered to be at the same level. For example, if the recommendation points corresponding to furniture and bed are added, because the level of furniture is higher than that of bed, and it has no relationship with food, rice, fruit, etc., it is initially determined to be the same as the level of food, but the bed is lower than the level of furniture, so furniture is at the first level and bed is at the second level; that is, food is compared and analyzed with rice, and the level is high, otherwise, the level of rice is low; based on the association and inclusion relationship of the recommended points, the corresponding judgment can be made through the existing recognition technology, so the association priority model can be established, which can be specifically established based on artificial intelligence technology and other methods; exemplary, the expression of the association priority model is:

[0062] ;

[0063] Where: (q, p) is the input data, q and p represent the two recommended points for comparative analysis; q→p means that the association level of q is higher than that of p; q←p means that the association level of q is lower than that of p; the output data is the association priority value GL(q, p), and the association priority values ​​are 0, 1, and -1;

[0064] When the association priority value is 0, it means that the two recommendation points have no association;

[0065] When the association priority value is 1, it means that the two recommended points are related, and the association level of q is higher than that of p;

[0066] When the association priority value is -1, it means that the two recommended points are associated, and the association level of q is lower than that of p.

[0067] The association relationship between each recommendation point is evaluated through the association priority model; that is, all recommendation points are evaluated in pairs, that is, each recommendation point is evaluated. For example, if there are three recommendation points ABC, then AB, BC, and AC need to be evaluated.

[0068] According to the correlation between each recommendation point, classification is performed to form several categories, which are marked as tree-like material categories. That is, each correlation level is first determined as a first-level recommendation point, and each first-level recommendation point is classified in turn to form each category. If a recommendation point is associated with multiple first-level corresponding recommendation points, it is copied and assigned to the corresponding category. The corresponding classification results can also be formed through other classification methods.

[0069] Each association tree is formed according to the association relationship between each recommendation point in each tree material category, that is, first distribute according to the association level, and then connect according to the association to form an association tree.

[0070] In one embodiment, a method for generating a recommended schematic diagram according to each association tree is as follows:

[0071] Identify the association levels corresponding to the association trees, and generate initial layers for the association trees according to the association levels, that is, the recommended points of the same management level are located in the same initial layer, such as furniture, bed, folding bed and food, rice, black rice, then furniture and food are located in the same layer, bed and rice are located in the same layer, and so on; identify the number of recommended points corresponding to each initial layer, and mark them as layer values;

[0072] The platform sets a display limit based on the best visualization display. That is, in order to facilitate display to users, no more than a certain number of recommendation points can be displayed at one time on the same interface. For example, if three-dimensional space is used to display the recommendation points, the display limit of the space is set, and the user needs to zoom in for further display. The initial layers are adjusted according to the display limit and the values ​​of each layer to obtain the display layers. Each display layer is visualized according to the preset display method to obtain the recommended schematic diagram.

[0073] In one embodiment, the method for adjusting each initial layer according to the display limit value and each layer value includes:

[0074] Determine the reference layer in each initial layer. The reference layer may be designated by the platform or the initial layer of the intermediate management level. The reference layer is determined according to the actual needs and display effects of the platform. The reference layer is the recommended points corresponding to the layer directly viewed by subsequent users. Identify the layer value corresponding to the reference layer and mark it as the reference value.

[0075] When the base value is not greater than the display limit, the base layer is not processed;

[0076] When the benchmark value is greater than the display limit, the benchmark layer is merged;

[0077] Sort and distribute the baseline layer and the remaining initial layers according to the correlation levels of the corresponding recommended points;

[0078] Compare the layer values ​​of the remaining initial layers with the display limit value, determine the initial layers that need to be merged, and merge the initial layers that need to be merged, that is, process them according to the merging method of the base layers;

[0079] Mark each initial layer and base layer as a display layer.

[0080] In one embodiment, the method for merging the reference layers includes:

[0081] Calculate the difference between the benchmark value and the display limit, and mark it as the display difference; perform simulated upgrading according to the display difference, and obtain various simulated upgrading methods, that is, use the recommended points of the higher management level to replace the recommended points of the next level corresponding to the recommended points in the benchmark layer, such as food-rice-black rice, yellow rice; black rice and yellow rice are of the same associated level, the benchmark layer is the layer corresponding to black rice and yellow rice, and the simulated upgrading is to replace black rice and yellow rice with rice of a higher level of black rice and yellow rice to reduce the benchmark value; determine various methods to eliminate the display difference according to the above method, and mark them as simulated upgrading methods, as long as the benchmark value is not greater than the display limit; each simulated upgrading method will not change without increasing or decreasing the recommended points, so each simulated upgrading method can be determined in advance, or it can be set by the staff of the platform;

[0082] The user's operation records are obtained in real time. If there are no user operation records, the representative operation records of the platform are used as representatives, such as the operation records of the users with the largest proportion, or designated by the platform. Matching and analyzing various simulated upgrade methods are performed according to the operation records, that is, simulation is performed according to each simulated upgrade method, and the corresponding adjusted baseline layer is determined. Preference matching is performed on each simulated adjusted baseline layer according to the operation records. The simulated upgrade method that can most clearly represent each recommended point in the operation record is used as the simulated upgrade method after the match is successful, and is marked as the target upgrade method, that is, it tends to upgrade each recommended point not included in the operation record and the recommended point with low weight determined according to the operation record; such as food-rice-black rice and furniture-bed-folding bed, because there is no furniture-bed-folding bed in the operation record, it tends to upgrade furniture-bed-folding bed;

[0083] The baseline layers are merged according to the target upgrading method. For example, black rice and yellow rice are replaced with rice of a higher grade than black rice and yellow rice, and black rice and yellow rice are combined into rice. The subsequent recommendation points for the rice can be enlarged to black rice and yellow rice.

[0084] Exemplarily, the method for matching and analyzing each simulated upgrade mode according to the operation record is as follows:

[0085] According to the operation records, the recommended points in the base layer and the record proportion of each recommended point are identified, and the record proportion is calculated according to the corresponding quantity;

[0086] The base layer is processed according to each simulation upgrade method, and the processed base layer is marked as the simulation layer; the record proportion of each recommended point in the simulation layer is marked as ηi, and for each upgraded recommended point, its record proportion is the sum of the corresponding merged record proportions; if there is no record proportion in the operation record, it is 0; i represents the subscript number of the recommended point in the simulation layer. If the simulation layer has three recommended points, then η1, η2, and η3 represent the record proportions of the above three recommended points respectively; i=1, 2, ..., n, where n is the number of recommended points in the simulation layer;

[0087] The simulation value of each simulation layer is calculated according to the simulation value formula; the simulation value formula is:

[0088] ;

[0089] Where: PA is the analog value;

[0090] The simulation upgrading method corresponding to the simulation layer with the largest simulation value is selected as the target upgrading method.

[0091] In one embodiment, each display layer is visualized according to a preset display method, that is, the platform presets a corresponding display method, such as a spatial display method, sets a space, and arranges each display layer according to the zoom-in and zoom-out operation functions. Each recommendation point is set in the space and distributed according to the distribution method of existing recommendation points. It can also be distributed randomly. Generally, those with similar attributes are distributed together. Subsequent users can intuitively view each recommendation point in the corresponding baseline layer, and can display all or part of the corresponding display layer through operations such as zooming in and out.

[0092] Step 2: Identify the user's operation data in real time, perform statistical analysis on the operation data according to the recommendation schematic, determine the recommended weight value of each recommended point in the recommendation schematic, and supplement the recommended weight value of each recommended point in the recommendation schematic; generate a recommendation detail table based on the recommended points currently displayed in the recommendation schematic, and the recommendation detail table is used to count the recommended weight values ​​of each recommended point in the display layer; mark the current recommendation schematic as a data display diagram, and display the data display diagram to the user in real time.

[0093] In one embodiment, the operation data is statistically analyzed according to the recommendation principle diagram. Statistical analysis can be performed based on existing technologies to determine the record proportion of each recommendation point, which is marked as a recommendation value for distinction. Subsequently, the recommendation proportion of each recommendation point in the corresponding display layer is calculated and marked as a recommendation weight value. The recommendation value can also be adjusted in combination with other parameters such as price, platform revenue, etc., and the recommendation weight value is determined based on the adjusted recommendation value.

[0094] In one embodiment, the method for performing statistical analysis on the operation data according to the recommended schematic diagram includes:

[0095] Determine the operation record of each recommendation point based on the operation data and the recommendation principle diagram, preset the unit time period, that is, perform data statistics in each unit time period, such as the common one day, one week, etc., which is set according to the needs of the platform, and can be adjusted according to the accumulated time, such as one day as the unit time period in the early stage, and two days, three days, etc. as the unit time period in the future; generate the record curve of each recommendation point based on the unit time period and each operation record, the horizontal axis of the record curve is the time representative axis corresponding to each unit time period, such as T=1, which represents the length of the first unit time period, and T=2 represents the length of the corresponding two unit time periods, which is determined according to the specific corresponding unit time period; the vertical axis is the number of operations counted in the corresponding unit time period, that is, the number of browsing, recording, etc., marked as unit value;

[0096] Set the time correction function. The time correction function is used to correct the unit values ​​at different times. It is set from the perspective of the representative decline of the operation data accumulated by the user. The unit values ​​of the current period are representative and do not need to be corrected. The more you go back, the lower the representativeness. The unit values ​​are gradually reduced until they are reduced to the lowest correction coefficient, such as 0, 0.1, 0.2, etc.; an example is e -0.005T ; Specifically, the decay effect over time can be simulated based on the existing exponential decay function until the platform's requirements are met. It can also be set manually by the platform. If the time effect is not considered, the time correction function is always 1;

[0097] The recorded curve of each recommended point is corrected according to the time correction function, that is, the correction coefficient of the corresponding time is multiplied by the unit value to adjust the recorded curve, the corrected recorded curve is marked as the unit curve, the unit curve is fitted, and the unit function is obtained;

[0098] The unit function of each recommended point in the same display layer is marked as DH(T)j, where T is the horizontal axis variable of the unit function, such as the common time t, independent variable x, etc., which can also be regarded as each recommended point in the same display interface; j represents the subscript number of the recommended point in the same display layer, which is the same as the representation form of i; j=1, 2, ..., m, where m is the number of recommended points in the display layer;

[0099] Calculate the recommended weight value of each recommended point according to the recommended weight formula; the recommended weight formula is:

[0100] ;

[0101] Where: PKj is the recommended weight value; when the user has no operation data, it is displayed according to the recommended weight value preset by the platform, that is, no calculation is required if there is no user operation data, and the denominator will not be zero.

[0102] In one embodiment, according to user needs, stage adjustments can be made based on the data display chart to form a stage display chart, that is, the user sets the stage time, which starts from the current time and moves forward for a period of time, such as the previous week, the previous month, etc., to remove the recommendation impact of previous operation data; the time correction function is adjusted according to the stage time, that is, all correction coefficients outside the stage time are adjusted to 0, and then calculations are performed, and a new data display chart is generated according to the adjusted time correction function. For the purpose of distinction, it is regarded as a stage display chart.

[0103] Step 3: Make real-time recommendations to users based on the data display graph, that is, when the recommendation weight value of each recommendation point is clear, make recommendations based on the existing recommendation algorithm; the user interactively adjusts the data display graph according to the recommendation result, that is, adjusts the weight coefficient of the corresponding recommendation point in the data display graph, such as increasing the recommendation weight value of a certain recommendation point, and adaptively adjusts the others according to the proportion of the increased recommendation weight value, and the same applies to other multiple adjustments, ensuring that the recommendation weight value manually adjusted by the user remains unchanged, and can be set to change with or without changes in the operation data;

[0104] The above formulas are all calculated by removing dimensions and taking numerical values. The formula is a formula that is closest to the actual situation obtained by collecting a large amount of data and performing software simulation. The preset parameters and preset thresholds in the formula are set by technical personnel in this field according to actual conditions or obtained by simulating a large amount of data.

[0105] The above embodiments are only used to illustrate the technical method of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical method of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present invention.

Claims

1. A visualization and human-computer interaction recommendation method for user operation data, characterized in that: Methods include: Determine each recommendation point in real time, perform association classification according to each recommendation point, and form each association tree; generate a recommendation principle diagram according to each association tree; Identify the user's operation data in real time, perform statistical analysis on the operation data according to the recommendation principle diagram, determine the recommendation weight value of each recommendation point in the recommendation principle diagram, and supplement the recommendation weight value of each recommendation point in the recommendation principle diagram; generate a corresponding recommendation detail table in real time according to the recommendation principle diagram; mark the current recommendation principle diagram as a data display diagram, and display the data display diagram to the user in real time; Making real-time recommendations to the user based on the data display graph, and the user interactively adjusting the data display graph based on the recommendation results; Methods for statistical analysis of operating data according to the recommended schematic diagram include: Determine the operation record of each recommended point according to the operation data and the recommended principle diagram, preset a unit period, and generate a record curve of each recommended point according to the unit period and each operation record, wherein the horizontal axis of the record curve is the time representative axis corresponding to each unit period; the vertical axis is the number of operations counted in the corresponding unit period, marked as a unit value; Set a time correction function; correct the recorded curve of each recommended point according to the time correction function, mark the corrected recorded curve as a unit curve, fit the unit curve, and obtain the unit function; The unit function of each recommended point in the same display layer is marked as DH(T)j, where T is the horizontal axis variable of the unit function, j=1, 2, ..., m, and m is the number of recommended points in the display layer; According to the formula Calculate the recommendation weight value of each recommendation point; Where: PKj is the recommended weight value; ; When the data display graph does not meet the user's needs, the user sets the stage time and adjusts the correction coefficient corresponding to the time correction function outside the stage time to 0; the adjusted time correction function is used to make corrections to form a new data display graph.

2. The method for visualization and human-computer interaction recommendation of user operation data according to claim 1, characterized in that: The method for association analysis based on each recommended point is: Establishing an association priority model, evaluating each pair of recommendation points through the association priority model, and obtaining the association relationship between each recommendation point; Classification is performed according to the association relationship between the recommendation points to form a plurality of tree-like material classes, the association relationship between the recommendation points in each tree-like material class is identified, and an association tree corresponding to each tree-like material class is formed according to each association relationship.

3. The method for visualization and human-computer interaction recommendation of user operation data according to claim 2, characterized in that: The expression of the association priority model is: ; Where: (q, p) is the input data, q and p represent the two recommended points for comparative analysis; q→p means that the correlation level of q is higher than that of p; q←p means that the correlation level of q is lower than that of p; The output data is the association priority value GL(q, p), and the association priority value is 0, 1, -1; When the association priority value is 0, it means that there is no association between the two recommendation points; When the association priority value is 1, it means that the two recommended points are related, and the association level of q is higher than that of p; When the association priority value is -1, it means that the two recommended points are associated, and the association level of q is lower than that of p.

4. The visualization and human-computer interaction recommendation method of user operation data according to claim 1, characterized in that: The method for generating the recommended schematic diagram based on each association tree is: Identify each association level corresponding to each association tree, generate each initial layer according to each association level, and the association level of each recommended point in the initial layer is the same; identify the number of recommended points corresponding to each initial layer, and mark them as layer values; The initial layers are adjusted according to preset display limits and layer values ​​to obtain display layers, and the display layers are visualized according to preset display methods to obtain a recommended schematic diagram.

5. The method for visualization and human-computer interaction recommendation of user operation data according to claim 4, characterized in that: Methods for adjusting each initial layer according to the display limit and each layer value include: Determine a reference layer in each of the initial layers, identify a layer value corresponding to the reference layer, and mark it as a reference value; When the base value is not greater than the display limit, the base layer is not processed; When the benchmark value is greater than the display limit, the benchmark layer is merged; Sort and distribute the baseline layer and the remaining initial layers according to the correlation levels of the corresponding recommended points; Compare the layer values ​​of the remaining initial layers with the display limit value, determine the initial layers that need to be merged, and merge the initial layers that need to be merged; Mark each initial layer and base layer as a display layer.

6. The method for visualization and human-computer interaction recommendation of user operation data according to claim 5, characterized in that: Methods for merging base layers include: Calculate the difference between the benchmark value and the display limit value, and mark it as the display difference; perform simulated upgrade according to the display difference, and obtain various simulated upgrade methods; Acquire the user's operation records in real time, perform matching analysis on each of the simulated upgrade methods according to the operation records, and determine the target upgrade method; The base layers are merged according to the target upgrading method.

7. The method for visualization and human-computer interaction recommendation of user operation data according to claim 6, characterized in that: The method for matching and analyzing each simulated upgrade method according to the operation records is as follows: Identify each recommended point in the reference layer and the record weight of each recommended point according to the operation record; Process the reference layer according to each of the simulation upgrading methods, and mark the processed reference layer as a simulation layer; mark the record weight of each recommended point in the simulation layer as ηi, where i=1, 2, ..., n, and n is the number of recommended points in the simulation layer; According to the formula Calculate the simulation value of each simulation layer; where: PA is the simulation value; The simulation upgrading method corresponding to the simulation layer with the largest simulation value is selected as the target upgrading method.

8. The method for visualization and human-computer interaction recommendation of user operation data according to claim 1, characterized in that: When users set the recommendation weight value of the recommendation point in the data display chart, they have two additional function options, one additional function option is dynamic update and fixed duration; the other additional function option is fixed.

Images