User intention-oriented data binning interaction method and device and storage medium

Abstract

The invention discloses a user intention-oriented data binning interaction method and device and a storage medium, and relates to the technical field of data binning, the method comprises the following steps: when a data binning node is in a to-be-interacted state, displaying an interaction page according to an interactive binning instruction; box separation adjustment parameters corresponding to the interactive page are determined, the box separation adjustment parameters are generated by the user through box separation adjustment according to the feature tag of the interactive page, and the interactive box separation instruction and the box separation adjustment parameters are generated with the user intention as the guide; and performing data binning according to the binning adjustment parameter and the binning index to obtain a binning boundary, and displaying a binning result to a user based on box monotonicity. Due to the fact that the user is guided to conduct interactive box separation through the interactive page and directly participates in the data box separation process, the box separation indexes and the feature tags of box separation are dynamically adjusted through interaction with the data box separation nodes, the intention of the user is fully played, and various complex and changeable data conditions and modeling requirements can be achieved.

Description

Technical Field

[0001] The present application relates to the technical field of data binning, and in particular to a method, device and storage medium for data binning interaction guided by user intention. Background Art

[0002] Data determines the upper limit of the effect of machine learning models, but different models have different needs for data, and different feature engineering, that is, data reorganization, is needed to better help the model learn the rules in the data and achieve good prediction results. Among them, data binning is to group the data and increase the number of data groups to prepare for more complex feature engineering later.

[0003] However, the data binning functions of various machine learning platforms are mostly fixed parameters. After the user has completed the data binning, they can no longer make interactive adjustments or simply interact with the results of the data binning, resulting in the subsequent modeling effect being completely dependent on the platform's built-in algorithms, with no room for user adjustment. As a result, the platform cannot cover the professional knowledge of each industry, does not reach the depth of actual development modeling data mining, and it is difficult to build a model that can be used in actual production. Summary of the invention

[0004] The main purpose of the present application is to provide a data binning interaction method, device and storage medium guided by user intention, aiming to solve the technical problem that most of the current data binning functions are fixed parameters or simple interactions, and the data binning results generated in this way have no room for user adjustment, making it difficult to achieve various complex and changeable data situations and modeling requirements.

[0005] To achieve the above objectives, the present application proposes a data binning interaction method guided by user intention, the method comprising:

[0006] When a data binning node is in a waiting-for-interaction state, an interactive page is displayed according to the user's interactive binning instruction, wherein the data binning node is a node in a visual modeling project of a machine learning platform;

[0007] Determine a bin adjustment parameter corresponding to the interactive page, wherein the bin adjustment parameter is generated by the user performing bin adjustment according to a feature tag of the interactive page, and the interactive bin instruction and the bin adjustment parameter are generated based on user intention;

[0008] Binning the data according to the binning adjustment parameter and the binning index selected by the user to obtain binning boundaries;

[0009] The binning result of the data binning node is displayed to the user according to the binning boundary and bin monotonicity.

[0010] In one embodiment, the step of determining the bin adjustment parameters corresponding to the interactive page, wherein the bin adjustment parameters are generated by the user performing bin adjustment according to the feature tags of the interactive page, includes:

[0011] Display a binning chart according to the feature switching instruction input by the user, the binning chart is used for the feature label of the bin selected by the user, and the binning chart includes a bin data volume chart, a user binning index chart, and a binning statistical table;

[0012] A bin adjustment operation is performed according to the feature label to obtain a bin adjustment parameter, wherein the bin adjustment operation includes merging bins and / or splitting bins.

[0013] In one embodiment, when the bin adjustment operation is bin merging, the step of performing the bin adjustment operation according to the feature tag to obtain the bin adjustment parameter includes:

[0014] Determine the data type of the feature tag, the data type includes a numeric type and a character type;

[0015] Merging the multiple boxes selected by the user according to the data type to obtain a merged new box, wherein the boxes are merged by box boundaries when they are numerical types, and merged by character values ​​when they are character types;

[0016] According to the merged new box, the box adjustment parameters are determined.

[0017] In one embodiment, when the bin adjustment operation is to split the bin, after the step of determining the data type of the feature label, the step further includes:

[0018] Determine the split point of the single box selected by the user, where the split point is the box boundary when the single box is a numerical value, and the split point is the character value when the single box is a character type;

[0019] Splitting the single box according to the data type and the split point to obtain multiple new boxes after splitting;

[0020] Based on the multiple new boxes after the splitting, a box adjustment parameter is determined.

[0021] In one embodiment, the step of performing data binning according to the binning adjustment parameter and the binning index selected by the user to obtain binning boundaries includes:

[0022] Displaying the user binning indicator graph to the user, wherein the user binning indicator graph is used to guide the user to select a binning indicator;

[0023] Binning the data according to the binning index and the data binning algorithm built into the machine learning platform, and obtaining multiple index results corresponding to the binning adjustment parameters;

[0024] The multiple indicator results are used as bin boundaries of the feature labels.

[0025] In one embodiment, the step of displaying the binning result of the data binning node to the user according to the binning boundary and bin monotonicity includes:

[0026] Sorting the bin boundaries according to bin monotonicity to obtain a sorted list;

[0027] Determine the binning result of the data binning node according to the application instruction generated by the user based on the sorted list, and display it to the user;

[0028] After the binning result is determined, the node state of the data binning node is changed to a completed state.

[0029] In addition, to achieve the above purpose, the present application also proposes a data binning interaction device guided by user intention, the device comprising:

[0030] An interactive page module, used to display an interactive page according to a user's interactive binning instruction when a data binning node is in a waiting-for-interaction state, wherein the data binning node is a node in a visual modeling project of a machine learning platform;

[0031] A bin adjustment module, used to determine bin adjustment parameters corresponding to the interactive page, wherein the bin adjustment parameters are generated by the user performing bin adjustment according to feature tags of the interactive page, and the interactive bin instructions and the bin adjustment parameters are generated based on user intentions;

[0032] A binning boundary module, used to bin data according to the binning adjustment parameter and the binning index selected by the user, and obtain binning boundaries;

[0033] The binning result module is used to display the binning result of the data binning node to the user according to the binning boundary and bin monotonicity.

[0034] In addition, to achieve the above-mentioned purpose, the present application also proposes a machine learning platform, which includes: a memory, a processor, and a computer program stored on the memory and executable on the processor, and the computer program is configured to implement the steps of the user intention-oriented data binning interaction method as described above.

[0035] In addition, to achieve the above-mentioned purpose, the present application also proposes a storage medium, which is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by the processor, the steps of the user intention-oriented data box interaction method as described above are implemented.

[0036] In addition, to achieve the above-mentioned purpose, the present application also provides a computer program product, which includes a computer program, and when the computer program is executed by a processor, it implements the steps of the user intention-oriented data binning interaction method as described above.

[0037] One or more technical solutions proposed in this application have at least the following technical effects: first, when the data binning node is in a waiting-for-interaction state, the application displays an interactive page according to the user's interactive binning instruction, and the data binning node is a node in the visual modeling project of the machine learning platform; then, the binning adjustment parameter corresponding to the interactive page is determined, and the binning adjustment parameter is generated by the user for binning adjustment according to the feature label of the interactive page, and the interactive binning instruction and the binning adjustment parameter are generated based on the user's intention; then, data binning is performed according to the binning adjustment parameter and the binning index selected by the user to obtain the binning boundary; finally, according to the binning boundary and the monotonicity of the binning body, the binning result of the data binning node is displayed to the user. Because the application guides users to interactively bin with their own professional knowledge through the interactive page, directly participates in the process of data binning, and dynamically adjusts the binning index and feature label of the binning by interacting with the data binning node, it avoids the situation of relying entirely on the platform's built-in algorithm, fully exerts the user's intention, and can realize various complex and changeable data situations and modeling requirements. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0039] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0040] Figure 1 A state change diagram of a data binning node in the user intent-oriented data binning interaction method of this application;

[0041] Figure 2 It is the interactive icon of the data binning node;

[0042] Figure 3 The completion icon for the data binning node;

[0043] Figure 4 A flowchart diagram of the first embodiment of the data binning interaction method guided by user intent of the present application is provided;

[0044] Figure 5 The box data volume diagram provided in Example 1;

[0045] Figure 6 The user binning indicator diagram provided in Example 1;

[0046] Figure 7 A schematic diagram of a binning statistics table provided in Example 1;

[0047] Figure 8 A flowchart diagram of the second embodiment of the data binning interaction method guided by user intent of the present application is provided;

[0048] Fig. 9 This is a schematic diagram of the module structure of a data binning interaction device guided by user intention in an embodiment of the present application;

[0049] Fig.10 Schematic diagram of the device structure of the hardware operating environment involved in the user intention-oriented data binning interaction method in the embodiment of the present application.

[0050] The realization of the purpose, functional features and advantages of this application will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0051] It should be understood that the specific embodiments described herein are only used to explain the technical solutions of the present application and are not used to limit the present application.

[0052] In order to better understand the technical solution of the present application, a detailed description will be given below in conjunction with the accompanying drawings and specific implementation methods.

[0053] It should be noted that the problem with the existing machine learning platforms is that they do not fully utilize user intentions, nor do they provide professional data binning prompts to guide users to use data binning technology with their own professional knowledge. As a result, there is no in-depth integration of users' industry experience and professional knowledge with AI modeling technology, and the depth of actual development modeling data mining has not been reached, making it difficult to build a model that can be used in actual production.

[0054] Therefore, this application proposes a new user-intention-oriented data binning interactive method that takes into account the binning effect to make up for the shortcomings of the above-mentioned problems. This application proposes interactive binning, which is a more flexible and interactive data binning method. In this mode, users can directly participate in the binning process and can dynamically adjust the binning rules, interval division, etc. by interacting with the data binning tool or system according to their own experience, knowledge and specific needs to achieve various complex and changeable data situations and modeling requirements.

[0055] For easier understanding, refer to Figure 1 , Figure 1 This is a state change diagram of the data binning node in the data binning interaction method guided by user intention in this application. The data binning interaction of this application is a node in the visual modeling project based on the machine learning platform, that is, the data binning node under feature preprocessing in the machine learning platform. During the execution of the data binning node, the user can check the interactive binning option, and the nodes in the visual modeling process will be executed in sequence until the data binning node icon changes to the state to be interacted (such as Figure 2 As shown, Figure 2 is the interactive icon of the data binning node. After the data set performs character mapping, the icon of the data binning node enters the interactive icon state). After the data binning node enters the interactive state, you can right-click the data binning and select to enter the interactive binning page. On the interactive binning page, the user can select the binning method of each feature and related parameters to make corresponding interactive binning adjustments. After the adjustment is completed, you can complete or exit the interactive binning page, and perform binning calculations to obtain the binning boundaries of each feature data. At this time, the node status is in execution. After the binning calculation is completed, the data binning node is completed. At this time, the node status is in execution, such as Figure 3 As shown, Figure 3 The completion icon of the data binning node is displayed, and the data binning node status changes to completion. After the data binning node is completed, the next node of the machine learning platform will be continued, which will not be introduced in detail here.

[0056] Through the user intention-oriented data binning interaction method of this application, the user intention can be fully utilized, and professional data binning prompts can be given to guide users to use data binning technology with their own professional knowledge, deeply integrate the user's industry experience, professional knowledge and modeling technology, achieve the depth of actual development modeling data mining, and facilitate the subsequent construction of models that can be used for actual production.

[0057] It should be noted that the execution subject of this embodiment can be a computing service device with data processing, interactive page display and data binning functions, such as a personal computer, server, etc., or an electronic device capable of realizing the above functions, a user-intention-oriented data binning interaction device (such as a machine learning platform) that executes the user-intention-oriented data binning interaction method of this application, etc., and this embodiment does not limit this. The following takes the machine learning platform as an example to illustrate this embodiment and the following embodiments.

[0058] Based on this, the embodiment of the present application provides a data binning interaction method guided by user intention, referring to Figure 1 and Figure 4 , Figure 4 A flowchart diagram of the first embodiment of the data binning interaction method guided by user intent of the present application is provided.

[0059] In this embodiment, the data binning interaction method guided by user intention includes steps S10 to S40:

[0060] Step S10: When the data binning node is in a waiting-for-interaction state, an interactive page is displayed according to the user's interactive binning instruction, and the data binning node is a node in a visual modeling project of a machine learning platform.

[0061] It should be noted that data binning is the process of dividing continuous numerical data or multiple discrete character data in a machine learning platform (a system that integrates a variety of machine learning related functions and tools for data preprocessing, model training and establishment, etc. Among them, the visual modeling project is a project that presents the complex process of the machine learning platform in an intuitive and easy-to-understand form through a graphical interface to facilitate the interaction of data binning) into a series of intervals or categories (i.e. boxes). The data binning node is the visualization node of the data binning process in the machine learning platform. Common data binning methods may include equal-width binning, equal-frequency binning, clustering-based binning, etc.

[0062] Through data binning, a large number of numerical values ​​can be converted into relatively less category information to reduce the data complexity of the machine learning platform, making it easier for the machine learning platform model to process and understand the data; and enhance model stability and reduce the impact of outliers on the model; at the same time, missing values ​​can also be processed.

[0063] It should be noted that the waiting for interaction state refers to the state in which the data binning node is waiting for the user to operate and interact. The interactive binning instruction is an instruction for the user to perform binning operations on the data binning node in an interactive manner. The interactive page is a page for the user to interact with the data binning node to adjust the data.

[0064] For ease of understanding, Figure 1As shown in the figure, if the user has checked the interactive binning option in the data binning node, after the data binning node is executed, the node status will change to the waiting for interaction status, and a status prompt will be given (for example, data binning has been completed, please enter the interactive page in the right-click menu of the node); at this time, the user can right-click the interactive option in the option (that is, generate interactive binning instructions) to interactively adjust the data binning results.

[0065] Furthermore, considering that the data binning node may have historical interactive data, the process of enabling interactive binning is as follows:

[0066] 1. The "interactive binning" configuration is enabled. Repeat the data binning node (i.e. check the interactive binning). Before running the interactive binning, if the user has enabled the "Whether to use interactive binning history" configuration, first determine whether there is a historical interactive binning configuration. If so, compare the current binning configuration. If there are fields with historical interactive binning configurations, retain the historical interactive binning configurations. If there are no historical configuration fields, use the data binning configuration. In this way, the interactive binning configuration of this task is combined to bin the data. If the user has not enabled the "Whether to use interactive binning history" configuration, use the data binning configuration for interactive binning.

[0067] 2. The "Interactive Binning" configuration is not enabled. If you run the data binning repeatedly (i.e., interactive binning is not enabled), the data binning configuration will be used to bin the data.

[0068] Through the above operations, when the user turns on the "interactive binning" and "whether to use interactive binning history" configurations on the machine learning platform, and then uses data that is exactly the same or partially the same as the historical binning data to repeatedly run the data binning node, there is no need to make repeated interactive adjustments to each feature, and the historical interactive binning configuration can be used. Of course, if the user is not satisfied with the previous interaction results, he or she can continue to make adjustments, and this embodiment does not limit this.

[0069] In this embodiment, the interactivity in the binning node means that users can participate in it and adjust the binning parameters, rules or methods in real time, so as to flexibly perform data binning processing according to specific needs and observations, thereby better analyzing and understanding the characteristics and distribution of the data.

[0070] Step S20: determining the bin adjustment parameters corresponding to the interactive page, wherein the bin adjustment parameters are generated by the user according to the feature tags of the interactive page, and the interactive bin instructions and the bin adjustment parameters are generated based on the user's intention.

[0071] It should be noted that the feature label is a classification label of the data feature for data binning in the interactive page.

[0072] It should be noted that the bin adjustment parameters are the adjusted parameters obtained after the user merges, splits, or performs other data binning operations on the feature labels (i.e., bins) in the interactive page, including the number of bins, the width or boundary value of each bin, and the binning rules (such as equidistant binning, equal-frequency binning, etc.).

[0073] Specifically, the bin adjustment process on the interactive page may include merging (such as upward merging and downward merging) and splitting, as follows:

[0074] 1. Merge. Continuously select multiple (more than two) cells (i.e. boxes) in the segment of the data feature in the interactive page, and then you can merge the selected boxes into one.

[0075] 2. Merge upwards. Select a single box, that is, select a single cell (not the first one) in the segment of the data feature in the interactive page, and then merge the selected cell (i.e. box) with the previous cell (i.e. box).

[0076] 3. Merge down. Select a single box, that is, select a single cell (not the last one) in the segment of the data feature in the interactive page, and then merge the selected cell (i.e. box) with the next cell (i.e. box).

[0077] 4. Split. Select a single box, that is, select a single cell in the segment of the data feature in the interactive page, and then split the selected cell into two cells (i.e. boxes) according to the prompt, that is, split one box into two boxes.

[0078] In this implementation, through the above-mentioned binning adjustment, the user's intention can be fully utilized, and the user can be guided to use data binning with his or her own professional knowledge.

[0079] In a feasible implementation, step S20 of this embodiment may include the steps of: displaying a binning chart according to the feature switching instruction input by the user, the binning chart being used for the feature label of the bin selected by the user, the binning chart including a bin data volume chart, a user binning indicator chart, and a binning statistical table; performing a binning adjustment operation according to the feature label to obtain binning adjustment parameters, the binning adjustment operation including merging bins and / or splitting bins.

[0080] It should be noted that the feature switching instruction is an instruction to switch various feature labels, switching one feature label to another feature label to achieve bin adjustment.

[0081] It should be noted that the binning chart is a chart used to display the binning process and adjust the binned data in the visual modeling project of the data binning node.

[0082] It should be noted that the box data volume chart is a chart used to display data volume information related to the box.

[0083] For ease of understanding, refer to Figure 5 , Figure 5 This is a box data volume graph provided in Example 1. As shown in the figure, the horizontal axis represents each bin (e.g., bin 0 to bin 9), and the vertical axis represents the bin data volume (e.g., 0 to 1200). After the user selects the data bin, the box data volume graph can be used to display the data volume of the sample in each bin.

[0084] It should be noted that the user binning indicator chart is an icon that displays the performance of related indicators in different binning intervals after the user bins (divides the intervals) the data features.

[0085] For ease of understanding, refer to Figure 6 , Figure 6 This is a user binning index graph provided in Example 1. As shown in the figure, the horizontal axis represents each bin (e.g., bin 0 to bin 9), and the vertical axis represents the index of the bin (e.g., 0 to 1). After the user selects the data bin, the index value of each bin can be displayed through the user binning index graph.

[0086] The index value can be determined by the weight value. The higher the weight value, the higher the index value, which is used to indicate the difference between the distribution of good samples and bad samples in the box. The calculation formula of the weight value is as follows:

[0087] WOE_i=ln(P_i / (1-P_i));

[0088] Among them, Pi represents the proportion of positive samples in a certain bin, 1-Pi represents the proportion of negative samples in a certain bin, ln represents the natural logarithm, and WOE represents the weight value.

[0089] Furthermore, the index value can also be determined by the bad rate, which is used to indicate the proportion of bad samples in the box. The calculation formula of the bad rate is: Bad_Rate (ie bad rate) = the number of bad samples in the box / the total number of samples in the box.

[0090] Furthermore, the index value can also be determined by IV value (Information Value), which is used to characterize the ability of each bin to distinguish the target variable. The IV value calculation formula is as follows:

[0091] Box IV value = WOE_i × (Pi-Pi');

[0092] Among them, Pi is the ratio of bad samples in the box to the total bad samples, and Pi' is the ratio of good samples in the box to the total good samples.

[0093] It should be noted that the binned statistical table is a table formed by dividing the data into different intervals (i.e. boxes) according to certain rules and then performing statistics and sorting.

[0094] In the bin statistics table, information such as the number of data contained in each bin, the proportion, and the summary value of related features can be displayed.

[0095] For ease of understanding, refer to Figure 7 , Figure 7 Schematic diagram of binning statistics table provided in Example 1. Field is the feature selected by the user; label is the bin number after binning the feature data selected by the user; segmentation, if it is a numerical feature, is the boundary of the bin; if it is a character feature, is the set of non-repeating values ​​in the bin; total proportion is the proportion of the amount of data in the bin to the total amount of data; Bad number is the number of bad samples in the bin; Bad proportion is the proportion of the amount of bad samples in the bin to the amount of data in the bin; Good number is the number of positive samples in the bin; Good proportion is the proportion of the amount of positive samples in the bin to the amount of data in the bin; binning index is represented by bad rate, and other indicators can also be selected, which is not limited in this embodiment.

[0096] It should be noted that merging boxes is an operation of merging boxes in the bin adjustment operation, and splitting boxes is an operation of splitting boxes in the bin adjustment operation.

[0097] In this embodiment, the interactive page displays a box data volume chart, a user box index chart, and a box statistics table for users to choose from. Users can select features through buttons and feature names on the page. After switching features, the corresponding chart is displayed to achieve box adjustment.

[0098] In a feasible implementation manner, in this embodiment, when the bin adjustment operation is to merge bins, the step of performing the bin adjustment operation according to the feature label to obtain the bin adjustment parameters includes: determining a data type of the feature label, where the data type includes a numeric type and a character type; merging the multiple bins selected by the user according to the data type to obtain a merged new bin, merging the bins by bin boundaries when the bins are numeric types, and merging by character values ​​when the bins are character types; and determining the bin adjustment parameters according to the merged new bins.

[0099] It should be noted that the data type is the type of field when operating on the box. There are two data types for fields, one is numeric type (data features represented and measured by numbers, such as age, height, weight, price, etc.), and the other is character type (data features composed of characters, such as name, gender, address, color, etc.).

[0100] In the optional box of the bin statistics table, users can select a bin (the serial number in the label column, representing the bin number) and implement bin adjustment operations through options such as merge, split, merge up, and merge down.

[0101] For ease of understanding, the merging process of numeric fields is as follows:

[0102] The upper bound of the first box among the selected boxes (i.e., the box boundary, which is determined by the following Figure 7 The values ​​at both ends of the segment in the table are used to determine the feature), and the lower bound of the last box is combined into a new boundary, that is, a new box. The indicators of each box are then recalculated to generate a related chart for the feature.

[0103] For example, if the user selects labels 0 and 1 (box 0 and box 1) of the field, the selected boxes 0 and 1 will be merged into one box, namely the new box 0, with the boundary consisting of the upper boundary of the original box 0 and the lower boundary of the original box 1. The subsequent box numbers will be regenerated in sequence, and the statistical indicators of the corresponding binning chart will be recalculated according to the data of the new box.

[0104] For ease of understanding, the merging process of character fields is as follows:

[0105] When character boxes are merged, the values ​​of multiple selected boxes (i.e., character values) can be aggregated and then deduplicated to form a new box, and then the relevant chart of the feature can be calculated.

[0106] For example, if the user selects labels 2 and 3 (i.e., boxes 2 and 3) of the field, the selected boxes 2 and 3 can be merged into one box, i.e., the new box 2. The value is as follows Figure 7 The values ​​of No. 2 and No. 3 before the segmented columns in the table are merged are summarized and then removed for duplicates to form character values. The numbers of subsequent boxes are regenerated in sequence, and the statistical indicators of the corresponding charts will also be recalculated according to the data of the new boxes.

[0107] In this implementation, bins of different field types can be merged through bin statistics tables.

[0108] In a feasible implementation manner, when the bin adjustment operation is to split the bin, after the step of determining the data type of the feature label, the step further includes: determining the split point of the single bin selected by the user, where the split point is the bin boundary when the single bin is of numerical type, and the split point is the character value when the single bin is of character type; splitting the single bin according to the data type and the split point to obtain multiple new bins after the split; and determining the bin adjustment parameters based on the multiple new bins after the split.

[0109] It should be noted that the split point is the key node when performing the box splitting operation, that is, the boundary value for dividing different boxes; it is used to clearly distinguish different parts or intervals.

[0110] When splitting boxes, fields also have two data types, one is numeric (data features represented and measured by numbers. For example, age, height, weight, price, etc.), and the other is character (data features composed of characters, such as name, gender, address, color, etc.).

[0111] In the bin statistics table, when the field is listed as a numerical feature, select the bin, then right-click and check the "Split" option button in the options. After applying, the indicators of each bin can be recalculated to generate relevant charts for the feature.

[0112] For example, if the user selects box No. 1 of the field column (the value of the label column is 1), right-click the mouse and select the "Split" option. In the interactive page, you can drag the value bar and select the split point of box No. 1 to complete the splitting of the value column features.

[0113] For example, the box boundary (segment column value) of box 1 (label column value is 1) of the original field (such as the age field) is [33.4, 41.1]. The split point 36.4 is now selected to split it into boxes 1 and 2. The subsequent box numbers are regenerated in sequence, and the statistical indicators of the corresponding charts will also be recalculated according to the data of the new boxes.

[0114] When splitting a numeric field, when the user splits a single box of a numeric feature (which can only be split into two boxes), the upper and lower bounds of the two boxes can be automatically generated, and the user only needs to fill in one of the values ​​(split point).

[0115] In the bin statistics table, when the segment is listed as a character feature, you can select it individually and then split it with the right click (only one box can be split into two boxes).

[0116] For example, if the user selects box 2 (label column value is 2) of a field (for example, a month field), the character values ​​of the box are 'nov', 'jun', 'may', 'jul', 'jan', 'aug'. In the interactive page, the checked character values ​​are one box, and the unchecked ones are another box, completing the splitting of the character column features. In box 2 of the original field, the selected character values ​​('nov', 'jun', 'may') are one box, which is box 2, and the unselected character values ​​(jul', 'jan', 'aug') are box 3. The numbers of the subsequent boxes are regenerated in order, and the statistical indicators of the corresponding charts will also be recalculated according to the data of the new boxes.

[0117] In this implementation, bin splitting of different field types can be achieved through bin statistical tables.

[0118] Step S30: binning the data according to the binning adjustment parameter and the binning index selected by the user to obtain binning boundaries.

[0119] It should be noted that the binning index is an index that characterizes the data in different binning intervals after the user bins (divides the intervals) the data features. The binning index can be represented by a user binning index graph, which will not be described in detail in this embodiment.

[0120] It should be noted that the binning boundaries are the adjusted box boundaries determined after data binning.

[0121] Specifically, the data binning may be calculated by an equal frequency binning method or an equal distance binning method, and this embodiment does not impose any limitation on this.

[0122] Step S40: displaying the binning result of the data binning node to the user according to the binning boundary and bin monotonicity.

[0123] It should be noted that box monotonicity refers to the monotonic change characteristics of the data within the box (interval or range), such as an increasing or decreasing trend.

[0124] Specifically, the bin boundaries may be sorted in ascending or descending order according to the monotonicity of the bins, and the bin boundaries may be recommended to the user according to the sorting result to determine the bin result.

[0125] In the technical solution provided in this embodiment, first, when the data binning node is in a waiting state for interaction, an interactive page is displayed according to the user's interactive binning instruction, so that the user can flexibly perform data binning processing according to specific needs and observations. Then the user adjusts the binning according to the feature label, such as merging (such as merging upward and merging downward) and splitting, giving full play to the user's intention, guiding the user to use the data binning with his own professional knowledge, and generating binning adjustment parameters. Then, data binning is performed according to the binning adjustment parameters and the binning index selected by the user to obtain the binning boundary. Finally, the binning boundary can be sorted in increasing or decreasing order according to the monotonicity of the bin body, and the binning boundary is recommended to the user according to the sorting result to determine the binning result. Since this embodiment guides the user to perform interactive binning with his own professional knowledge through the interactive page, directly participates in the process of data binning, and dynamically adjusts the binning index and feature label of the binning by interacting with the data binning node, avoiding the situation of relying entirely on the platform's built-in algorithm, giving full play to the user's intention, and realizing various complex and changeable data situations and modeling requirements.

[0126] Based on the first embodiment of the present application, in the second embodiment of the present application, the same or similar contents as those in the above-mentioned embodiment 1 can be referred to the above introduction, and will not be repeated in the following. Figure 4 and Figure 8 , Figure 8 A flowchart diagram of the second embodiment of the data binning interaction method based on user intent provided in the present application.

[0127] In this example, step S30 includes steps S31 to S33:

[0128] Step S31: showing the user binning indicator graph to the user, wherein the user binning indicator graph is used to guide the user to select a binning indicator.

[0129] Step S32: binning the data according to the binning index and the data binning algorithm built into the machine learning platform to obtain multiple index results corresponding to the binning adjustment parameters.

[0130] It should be noted that the data binning algorithm divides continuous data into several intervals (i.e. "bins"). For example, binning methods such as equal frequency, equal interval, and decision tree can transform complex data structures into relatively simpler and more regular forms.

[0131] Step S33: using the multiple indicator results as bin boundaries of the feature labels.

[0132] In this implementation, the user can select a binning index before each binning adjustment. Based on the calculation principle of the binning index, the platform can try different built-in data binning methods (such as equal frequency, equal interval, decision tree and other binning methods) to bin the data, so as to calculate the binning index results under each binning adjustment parameter. Through the selected binning index, the data binning boundary recommendation is realized based on the user's intention.

[0133] Furthermore, step S40 of this example includes the steps of: sorting the binning boundaries according to the monotonicity of the bins to obtain a sorted list; displaying the binning results of the data binning nodes to the user according to the application instructions generated by the user based on the sorted list; and after the binning results are determined, changing the node state of the data binning nodes to a completed state.

[0134] It should be noted that the sort list is a list obtained by sorting the bin boundaries in ascending or descending order. The application instruction is an operation in which the user selects a bin boundary to apply according to the sort list.

[0135] Specifically, the platform can sort the bin boundaries of all bin adjustment parameter solutions according to bin monotonicity and recommend them to users. The user clicks the "Apply" button after the recommended result to generate an application instruction. The platform applies the bin result to the bin of the feature according to the application instruction.

[0136] In this implementation, data binning boundaries are recommended based on user intent, i.e., user binning indicators, to guide users to use data binning technology with their own professional knowledge, thereby deeply integrating users' industry experience and professional knowledge with AI modeling technology to achieve the depth of data mining during actual development modeling.

[0137] Furthermore, if the binning needs to be reset, the user can reset the selected features to the initial binning results through the interactive page.

[0138] In the technical solution provided in this embodiment, data binning boundaries are recommended based on user intention, that is, user binning indicators, to guide users to use data binning technology with their own professional knowledge, thereby deeply integrating users' industry experience and professional knowledge with AI modeling technology to achieve the depth of data mining during actual development modeling.

[0139] It should be noted that the above examples are only used to understand the present application and do not constitute a limitation on the user intention-oriented data binning interaction method of the present application. More simple transformations based on this technical concept are all within the scope of protection of the present application.

[0140] This application also provides a data binning interactive device guided by user intention, please refer to Fig. 9 , Fig. 9 This is a schematic diagram of the module structure of a data binning interaction device guided by user intention in an embodiment of the present application; the data binning interaction device guided by user intention includes:

[0141] An interactive page module 901 is used to display an interactive page according to a user's interactive binning instruction when a data binning node is in a waiting-for-interaction state, wherein the data binning node is a node in a visual modeling project of a machine learning platform;

[0142] A bin adjustment module 902, for determining bin adjustment parameters corresponding to the interactive page, wherein the bin adjustment parameters are generated by the user performing bin adjustment according to feature tags of the interactive page, and the interactive bin instruction and the bin adjustment parameters are generated based on user intention;

[0143] A binning boundary module 903 is used to bin data according to the binning adjustment parameter and the binning index selected by the user to obtain a binning boundary;

[0144] The binning result module 904 is used to display the binning result of the data binning node to the user according to the binning boundary and bin monotonicity.

[0145] In this embodiment, the bin adjustment module 902 is also used to display a bin chart according to the feature switching instruction input by the user, and the bin chart is used for the feature label of the bin selected by the user, and the bin chart includes a bin data volume chart, a user bin index chart and a bin statistics table; the bin adjustment operation is performed according to the feature label to obtain the bin adjustment parameters, and the bin adjustment operation includes merging bins and / or splitting bins.

[0146] Furthermore, when the bin adjustment operation is to merge bins, the bin adjustment module 902 is further used to determine the data type of the feature label, which data type includes a numeric type and a character type; merge the multiple bins selected by the user according to the data type to obtain a merged new bin, and merge the bins by bin boundaries when the bins are numeric types, and merge the bins by character values ​​when the bins are character types; and determine the bin adjustment parameters according to the merged new bins.

[0147] Furthermore, when the bin adjustment operation is to split the bin, the bin adjustment module 902 is further used to determine the split point of the single bin selected by the user, the split point is the bin boundary when the single bin is of numerical type, and the split point is the character value when the single bin is of character type; the single bin is split according to the data type and the split point to obtain multiple new bins after the split; and the bin adjustment parameters are determined based on the multiple new bins after the split.

[0148] Furthermore, the binning boundary module 903 is also used to display the user binning indicator graph to the user, and the user binning indicator graph is used to guide the user to select the binning indicator; binning the data according to the binning indicator and the data binning algorithm built into the machine learning platform to obtain multiple indicator results corresponding to the binning adjustment parameters; and use the multiple indicator results as the binning boundaries of the feature labels.

[0149] Furthermore, the binning result module 904 is also used to sort the binning boundaries according to the monotonicity of the bins to obtain a sorted list; determine the binning results of the data binning nodes according to the application instructions generated by the user based on the sorted list, and display them to the user; after the binning results are determined, change the node state of the data binning node to a completed state.

[0150] Other embodiments or specific implementations of the user intention-oriented data binning interaction device of the present application can refer to the above-mentioned method embodiments and will not be repeated here.

[0151] The user intention-oriented data binning interaction device provided by the present application adopts the user intention-oriented data binning interaction method in the above-mentioned embodiment, which can solve the technical problem that most of the current data binning functions are fixed parameters or simple interactions, and the data binning results generated in this way have no room for user adjustment, making it difficult to achieve various complex and changeable data situations and modeling requirements. Compared with the prior art, the beneficial effects of the user intention-oriented data binning interaction device provided by the present application are the same as the beneficial effects of the user intention-oriented data binning interaction method provided by the above-mentioned embodiment, and the other technical features in the user intention-oriented data binning interaction device are the same as the features disclosed in the above-mentioned embodiment method, which will not be repeated here.

[0152] The present application provides a machine learning platform, which includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the user intent-oriented data binning interaction method in the above-mentioned embodiment 1.

[0153] Reference below Fig.10 , Fig.10 The device structure diagram of the hardware operating environment involved in the user intention-oriented data binning interaction method in the embodiment of the present application shows a structural diagram of a machine learning platform suitable for implementing the embodiment of the present application. The machine learning platform in the embodiment of the present application may include but is not limited to mobile terminals such as mobile phones, laptops, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Descriptions), PMPs (Portable Media Players), vehicle-mounted terminals (such as vehicle-mounted navigation terminals), etc., and fixed terminals such as digital TVs, desktop computers, etc. Fig.10 The machine learning platform shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of the present application.

[0154] like Fig.10As shown, the machine learning platform may include a processing device 1001 (e.g., a central processing unit, a graphics processor, etc.), which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM: Read Only Memory) 1002 or a program loaded from a storage device 1003 to a random access memory (RAM: Random Access Memory) 1004. In RAM1004, various programs and data required for the operation of the machine learning platform are also stored. The processing device 1001, ROM1002, and RAM1004 are connected to each other via a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus. Typically, the following systems can be connected to the I / O interface 1006: an input device 1007 including, for example, a touch screen, a touchpad, a keyboard, a mouse, an image sensor, a microphone, an accelerometer, a gyroscope, etc.; an output device 1008 including, for example, a liquid crystal display (LCD: Liquid Crystal Display), a speaker, a vibrator, etc.; a storage device 1003 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 1009. The communication device 1009 can allow the machine learning platform to communicate with other devices wirelessly or by wire to exchange data. Although the figure shows a machine learning platform with various systems, it should be understood that it is not required to implement or have all the systems shown. More or fewer systems can be implemented or have alternatively.

[0155] In particular, according to the embodiments disclosed in the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, the embodiments disclosed in the present application include a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through a communication device, or installed from a storage device 1003, or installed from a ROM 1002. When the computer program is executed by the processing device 1001, the above-mentioned functions defined in the method of the embodiment disclosed in the present application are executed.

[0156] The machine learning platform provided by the present application adopts the data binning interaction method guided by user intention in the above-mentioned embodiment, which can solve the technical problem that most of the current data binning functions are fixed parameters or simple interactions, and the data binning results generated in this way have no room for user adjustment, making it difficult to achieve various complex and changeable data situations and modeling requirements. Compared with the prior art, the beneficial effects of the machine learning platform provided by the present application are the same as the beneficial effects of the data binning interaction method guided by user intention provided in the above-mentioned embodiment, and the other technical features in the machine learning platform are the same as the features disclosed in the method of the previous embodiment, which will not be repeated here.

[0157] It should be understood that the various parts disclosed in this application can be implemented by hardware, software, firmware or a combination thereof. In the description of the above embodiments, specific features, structures, materials or characteristics can be combined in any one or more embodiments or examples in a suitable manner.

[0158] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0159] The present application provides a computer-readable storage medium having computer-readable program instructions (ie, computer programs) stored thereon, wherein the computer-readable program instructions are used to execute the user intention-oriented data binning interaction method in the above-mentioned embodiment.

[0160] The computer-readable storage medium provided in the present application may be, for example, a USB flash drive, but is not limited to electrical, magnetic, optical, electromagnetic, infrared or semiconductor systems, systems or devices, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In this embodiment, the computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, system or device. The program code contained on the computer-readable storage medium may be transmitted using any appropriate medium, including but not limited to: wires, optical cables, RF (Radio Frequency), etc., or any suitable combination of the above.

[0161] The above-mentioned computer-readable storage medium may be included in the machine learning platform; or it may exist independently without being assembled into the machine learning platform.

[0162] The computer-readable storage medium carries one or more programs. When the one or more programs are executed by the machine learning platform, the machine learning platform: when a data binning node is in a waiting-for-interaction state, displays an interactive page according to the user's interactive binning instructions, and the data binning node is a node in a visual modeling project of the machine learning platform; determines the binning adjustment parameters corresponding to the interactive page, and the binning adjustment parameters are generated by the user for binning adjustment according to the feature labels of the interactive page, and the interactive binning instructions and the binning adjustment parameters are generated based on the user's intention; bins the data according to the binning adjustment parameters and the binning indicators selected by the user to obtain binning boundaries; and displays the binning results of the data binning node to the user according to the binning boundaries and the monotonicity of the bins.

[0163] Computer program code for performing the operations of the present application may be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, C++, and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0164] The flowchart and block diagram in the accompanying drawings illustrate the possible architecture, functions and operations of the system, method and computer program product according to various embodiments of the present application. In this regard, each box in the flowchart or block diagram can represent a module, a program segment or a part of a code, and the module, a program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flowchart and the combination of boxes in the block diagram and / or flowchart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0165] The modules involved in the embodiments described in the present application may be implemented by software or hardware, wherein the name of the module does not constitute a limitation on the unit itself in some cases.

[0166] The readable storage medium provided by the present application is a computer-readable storage medium, which stores computer-readable program instructions (i.e., computer programs) for executing the above-mentioned data binning interaction method guided by user intention, which can solve the technical problem that most of the current data binning functions are fixed parameters or simple interactions, and the data binning results generated in this way have no user-adjustable space, making it difficult to achieve various complex and changeable data situations and modeling requirements. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided by the present application are the same as the beneficial effects of the data binning interaction method guided by user intention provided by the above-mentioned embodiment, and will not be repeated here.

[0167] The present application also provides a computer program product, including a computer program, which, when executed by a processor, implements the steps of the above-mentioned user intention-oriented data binning interaction method.

[0168] The computer program product provided by the present application can solve the technical problem that most of the current data binning functions are fixed parameters or simple interactions, and the data binning results generated in this way have no room for user adjustment, making it difficult to achieve various complex and changeable data situations and modeling requirements. Compared with the prior art, the beneficial effects of the computer program product provided by the present application are the same as the beneficial effects of the user intention-oriented data binning interaction method provided in the above embodiment, and will not be repeated here.

[0169] The above descriptions are only some embodiments of the present application, and are not intended to limit the patent scope of the present application. All equivalent structural changes made using the contents of the present application specification and drawings under the technical concept of the present application, or direct / indirect applications in other related technical fields are included in the patent protection scope of the present application.

Claims

1. A data binning interaction method guided by user intention, characterized in that: The method includes: When a data binning node is in a waiting-for-interaction state, an interactive page is displayed according to the user's interactive binning instruction, wherein the data binning node is a node in a visual modeling project of a machine learning platform; Determine a bin adjustment parameter corresponding to the interactive page, wherein the bin adjustment parameter is generated by the user performing bin adjustment according to a feature tag of the interactive page, and the interactive bin instruction and the bin adjustment parameter are generated based on user intention; Binning the data according to the binning adjustment parameter and the binning index selected by the user to obtain binning boundaries; The binning result of the data binning node is displayed to the user according to the binning boundary and bin monotonicity.

2. The method according to claim 1, characterized in that The step of determining the bin adjustment parameters corresponding to the interactive page, wherein the bin adjustment parameters are generated by the user performing bin adjustment according to the feature tags of the interactive page, includes: Display a binning chart according to the feature switching instruction input by the user, the binning chart is used for the feature label of the bin selected by the user, and the binning chart includes a bin data volume chart, a user binning index chart, and a binning statistical table; A bin adjustment operation is performed according to the feature label to obtain a bin adjustment parameter, wherein the bin adjustment operation includes merging bins and / or splitting bins.

3. The method according to claim 2, characterized in that When the bin adjustment operation is bin merging, the step of performing the bin adjustment operation according to the feature tag to obtain the bin adjustment parameter includes: Determine the data type of the feature tag, the data type includes a numeric type and a character type; Merging the multiple boxes selected by the user according to the data type to obtain a merged new box, wherein the boxes are merged by box boundaries when they are numerical types, and merged by character values ​​when they are character types; According to the merged new box, the box adjustment parameters are determined.

4. The method according to claim 3, characterized in that When the bin adjustment operation is to split the bins, after the step of determining the data type of the feature label, the step further includes: Determine the split point of the single box selected by the user, where the split point is the box boundary when the single box is a numerical value, and the split point is the character value when the single box is a character type; Splitting the single box according to the data type and the split point to obtain multiple new boxes after splitting; Based on the multiple new boxes after the splitting, a box adjustment parameter is determined.

5. The method according to any one of claims 2 to 4, characterized in that The step of performing data binning according to the binning adjustment parameter and the binning index selected by the user to obtain binning boundaries includes: Displaying the user binning indicator graph to the user, wherein the user binning indicator graph is used to guide the user to select a binning indicator; Binning the data according to the binning index and the data binning algorithm built into the machine learning platform, and obtaining multiple index results corresponding to the binning adjustment parameters; The multiple indicator results are used as bin boundaries of the feature labels.

6. The method according to any one of claims 1 to 4, characterized in that The step of displaying the binning result of the data binning node to the user according to the binning boundary and bin monotonicity comprises: Sorting the bin boundaries according to bin monotonicity to obtain a sorted list; Determine the binning result of the data binning node according to the application instruction generated by the user based on the sorted list, and display it to the user; After the binning result is determined, the node state of the data binning node is changed to a completed state.

7. A data binning interactive device guided by user intention, characterized in that: The device comprises: An interactive page module, used to display an interactive page according to a user's interactive binning instruction when a data binning node is in a waiting-for-interaction state, wherein the data binning node is a node in a visual modeling project of a machine learning platform; A bin adjustment module, used to determine bin adjustment parameters corresponding to the interactive page, wherein the bin adjustment parameters are generated by the user performing bin adjustment according to feature tags of the interactive page, and the interactive bin instructions and the bin adjustment parameters are generated based on user intentions; A binning boundary module, used to bin data according to the binning adjustment parameter and the binning index selected by the user, and obtain binning boundaries; The binning result module is used to display the binning result of the data binning node to the user according to the binning boundary and bin monotonicity.

8. A machine learning platform, characterized in that: The machine learning platform includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program is configured to implement the steps of the user intent-oriented data binning interaction method as described in any one of claims 1 to 6.

9. A storage medium, characterized in that: The storage medium is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by a processor, the steps of the user intention-oriented data binning interaction method according to any one of claims 1 to 6 are implemented.

10. A computer program product, characterized in that The computer program product comprises a computer program, and when the computer program is executed by a processor, the steps of the user intention-oriented data binning interaction method according to any one of claims 1 to 6 are implemented.

Images