Attribute information determination method and device, equipment and computer storage medium

Automatically determine the user satisfaction evaluation attribute information through comprehensive weights and correlation accuracy, the problems of high labor costs and high subjectivity in the prior art are solved, and the accuracy and efficiency of satisfaction analysis are improved.

CN118941140APending Publication Date: 2024-11-12CHINA MOBILE INFORMATION TECHNOLOGY CO LTD +1
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410968831.1
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-18
Publication Date
2024-11-12

AI Technical Summary

Technical Problem

In the prior art, determining the key attribute information of user satisfaction evaluation information based on manual methods is high cost and has a large subjectivity, resulting in low accuracy.

Method used

Through comprehensive weights and correlation accuracy, select attribute information with a larger weight from the satisfaction evaluation information, use preset relationships to determine the target weight, and automatically determine the key satisfaction evaluation attribute information to reduce manual intervention.

Benefits of technology

It saves labor costs, improves the accuracy of key satisfaction evaluation attribute information, and realizes automated user satisfaction analysis.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118941140A_ABST
    Figure CN118941140A_ABST
Patent Text Reader

Abstract

The invention discloses an attribute information determination method and device, equipment and a computer storage medium. The method comprises the steps that multiple pieces of satisfaction evaluation information in a target time period and the weight of satisfaction evaluation attribute information in each piece of satisfaction evaluation information are acquired; selecting m pieces of target satisfaction degree evaluation attribute information with relatively large weights from the satisfaction degree evaluation attribute information, wherein m is a positive integer; obtaining the association accuracy of each piece of target satisfaction evaluation attribute information; determining a target weight of each piece of target satisfaction evaluation attribute information according to a preset relationship according to the weight and the association accuracy of each piece of target satisfaction evaluation attribute information; and determining a preset number of target satisfaction evaluation attribute information with relatively large target weights as key satisfaction evaluation attribute information. The key satisfaction evaluation attribute information is selected by integrating the weight and the association accuracy without depending on a manual mode, so that the labor cost is saved, and the accuracy of determining the key satisfaction evaluation attribute information is improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application belongs to the field of artificial intelligence technology, and in particular, relates to a method, device, equipment and computer storage medium for determining attribute information. Background Art

[0002] Operators need to consider user satisfaction while paying attention to the growth of the number of users. If user growth is based on sacrificing user usage perception, giving users a poor service experience is not conducive to the long-term development of the business. Therefore, the existing service operation system pays more and more attention to improving user satisfaction. In order to improve user satisfaction, it is necessary to analyze user satisfaction evaluation information.

[0003] In the prior art, the acquired satisfaction evaluation information is usually studied manually, and key satisfaction evaluation attribute information is extracted from the satisfaction evaluation information to determine the factors affecting user satisfaction. This method of determining key satisfaction evaluation attribute information not only has a relatively high labor cost, but also has a large degree of subjectivity, and the key satisfaction evaluation attribute information is not accurately determined. Summary of the invention

[0004] The embodiments of the present application provide a method, apparatus, device and computer storage medium for determining attribute information, which selects key satisfaction evaluation attribute information based on comprehensive weight and correlation accuracy, and no longer relies on manual methods, thereby saving labor costs and improving the accuracy of determining key satisfaction evaluation attribute information.

[0005] In a first aspect, an embodiment of the present application provides a method for determining attribute information, including:

[0006] Obtaining multiple satisfaction evaluation information in a target time period and the weight of satisfaction evaluation attribute information in each satisfaction evaluation information;

[0007] Select m target satisfaction evaluation attribute information with larger weights from the satisfaction evaluation attribute information, where m is a positive integer;

[0008] Obtain the correlation accuracy of each target satisfaction evaluation attribute information;

[0009] According to the weight and the associated accuracy of each target satisfaction evaluation attribute information, the target weight of each target satisfaction evaluation attribute information is determined according to the preset relationship, the preset relationship includes the relationship between the weight, the associated accuracy and the target weight of the preset satisfaction evaluation attribute information, and the preset satisfaction evaluation attribute information includes m target satisfaction evaluation attribute information;

[0010] A preset number of target satisfaction evaluation attribute information with a larger target weight is determined as key satisfaction evaluation attribute information.

[0011] In a possible implementation, the weight includes a first sub-weight, a second sub-weight and a third sub-weight; m target satisfaction evaluation attribute information with larger weights are selected from the satisfaction evaluation attribute information, including:

[0012] From the satisfaction evaluation attribute information, select k first satisfaction evaluation attribute information with larger weights in the first sub-weight, l second satisfaction evaluation attribute information with larger weights in the second sub-weight, and n third satisfaction evaluation attribute information with larger weights in the third sub-weight;

[0013] Merging the first satisfaction evaluation attribute information, the second satisfaction evaluation attribute information and the third satisfaction evaluation attribute information to obtain m target satisfaction evaluation attribute information;

[0014] Among them, the sum of the first sub-weights of the k first satisfaction evaluation attribute information, the sum of the second sub-weights of the l second satisfaction evaluation attribute information, and the sum of the third sub-weights of the n third satisfaction evaluation attribute information are all not less than the first threshold; m≤k+l+n.

[0015] In a possible implementation, obtaining the weight of the satisfaction evaluation attribute information in each satisfaction evaluation information includes:

[0016] Obtain the frequency of the satisfaction evaluation attribute information in each satisfaction evaluation information, the evaluation level of the satisfaction evaluation attribute information in each satisfaction evaluation information, and the probability of the satisfaction evaluation attribute information in the satisfaction evaluation information of each satisfaction level;

[0017] According to the frequency of the satisfaction evaluation attribute information, determine the proportion of the satisfaction evaluation attribute information in each satisfaction evaluation information;

[0018] According to the proportion of the satisfaction evaluation attribute information in each satisfaction evaluation information, the entropy of the satisfaction evaluation attribute information is determined;

[0019] Determine a first sub-weight of the satisfaction evaluation attribute information according to the entropy of the satisfaction evaluation attribute information;

[0020] determining an importance index of the satisfaction evaluation attribute information according to the evaluation level of the satisfaction evaluation attribute information in each satisfaction evaluation information and the number of the plurality of satisfaction evaluation attribute information;

[0021] Normalizing the importance index of the satisfaction evaluation attribute information to obtain the second sub-weight of the satisfaction evaluation attribute information;

[0022] The third sub-weight of the satisfaction evaluation attribute information is determined according to the probability of the satisfaction evaluation attribute information in the satisfaction evaluation information of each satisfaction level.

[0023] In a possible implementation, before obtaining the correlation accuracy of each target satisfaction evaluation attribute information, the method further includes:

[0024] Randomly select multiple first target satisfaction evaluation attribute information from the m target satisfaction evaluation attribute information;

[0025] For each selection of multiple first target satisfaction evaluation attribute information, a first satisfaction prediction model is constructed using the multiple first target satisfaction evaluation attribute information respectively, and the multiple satisfaction evaluation information is input into the first satisfaction prediction model, and the satisfaction levels corresponding to the multiple satisfaction evaluation information are predicted using the first satisfaction classification rules corresponding to the multiple first target satisfaction evaluation attribute information in the first satisfaction prediction model, so as to obtain predicted satisfaction levels corresponding to the multiple satisfaction evaluation information;

[0026] Determining the accuracy of the first satisfaction prediction model according to the predicted satisfaction levels and the actual satisfaction levels corresponding to the plurality of satisfaction evaluation information;

[0027] Calculate the average of the accuracy of the first satisfaction prediction model to obtain an average accuracy;

[0028] According to the average accuracy and the amount of the first target satisfaction evaluation attribute information, the associated accuracy of each target satisfaction evaluation attribute information is determined.

[0029] In a possible implementation, before obtaining the weight of the satisfaction evaluation attribute information in each satisfaction evaluation information, the method further includes:

[0030] Segment multiple satisfaction evaluation information to obtain multiple target segmentations;

[0031] Deleting the segmentations that are consistent with the preset stop words from the multiple target segmentations to obtain multiple first segmentations;

[0032] The segmentations in the first segmentations whose occurrence frequency in the plurality of satisfaction evaluation information is less than the second threshold are deleted to obtain the satisfaction evaluation attribute information.

[0033] In a possible implementation, the segmentations in the first segmentation whose occurrence frequency in the plurality of satisfaction evaluation information is less than the second threshold are deleted to obtain the satisfaction evaluation attribute information, including:

[0034] Deleting the first participles whose occurrence frequency in the plurality of satisfaction evaluation information is less than the second threshold, to obtain a plurality of second participles;

[0035] For each target second participle in the plurality of second participles, calculating the similarity between the target second participle and a third participle, where the third participle is a participle in the plurality of second participles except the target second participle;

[0036] From the plurality of second participles, third participles whose similarity to the target second participle is greater than a third threshold are deleted to obtain satisfaction evaluation attribute information.

[0037] In a possible implementation, third participles whose similarity to the target second participle is greater than a third threshold are deleted from the plurality of second participles to obtain satisfaction evaluation attribute information, including:

[0038] From the plurality of second participles, delete the third participles whose similarity to the target second participle is greater than a third threshold, to obtain a plurality of fourth participles;

[0039] A fourth participle whose appearance frequency in a plurality of preset words is greater than a fourth threshold is selected to obtain satisfaction evaluation attribute information.

[0040] In one possible embodiment, the method further includes:

[0041] Compare each satisfaction evaluation information with the satisfaction evaluation attribute information to obtain a comparison result;

[0042] According to the comparison results, characteristic information of each satisfaction evaluation information is generated;

[0043] The characteristic information is input into the second satisfaction prediction model, and the satisfaction levels corresponding to the multiple satisfaction evaluation information are predicted according to the second satisfaction classification rule in the second satisfaction prediction model to obtain the satisfaction levels corresponding to the multiple satisfaction evaluation information.

[0044] In a possible implementation, before inputting the feature information into the second satisfaction prediction model, the method further includes:

[0045] Obtaining multiple satisfaction evaluation information samples and actual satisfaction levels corresponding to the satisfaction evaluation information samples;

[0046] Segmenting multiple satisfaction evaluation information samples to obtain multiple satisfaction evaluation attribute information samples;

[0047] Compare each satisfaction evaluation information sample with multiple satisfaction evaluation attribute information samples to obtain a sample comparison result;

[0048] According to the sample comparison results, a characteristic information sample of each satisfaction evaluation information is generated;

[0049] Inputting the characteristic information samples into the initial second satisfaction prediction model, predicting the satisfaction levels corresponding to the multiple satisfaction evaluation information samples according to the initial second satisfaction classification rules in the initial second satisfaction prediction model, and obtaining the sample predicted satisfaction levels corresponding to the multiple satisfaction evaluation information samples;

[0050] Determining the accuracy of the initial second satisfaction prediction model according to the sample prediction satisfaction levels corresponding to the multiple satisfaction evaluation information samples and the actual satisfaction levels of the multiple satisfaction evaluation information samples;

[0051] Update the initial second satisfaction classification rules in the initial second satisfaction prediction model, predict the satisfaction levels corresponding to multiple satisfaction evaluation information samples according to the updated initial second satisfaction classification rules, and obtain updated sample predicted satisfaction levels; update the accuracy of the initial second satisfaction prediction model according to the updated sample predicted satisfaction levels and the actual satisfaction levels of the multiple satisfaction evaluation information samples; when the accuracy of the initial second satisfaction prediction model is greater than the fourth threshold, obtain the second satisfaction prediction model, and the second satisfaction prediction model includes the second satisfaction classification rules.

[0052] In a second aspect, an embodiment of the present application provides a device for determining attribute information, including:

[0053] An acquisition module, used to acquire multiple satisfaction evaluation information of a target time period and the weight of satisfaction evaluation attribute information in each satisfaction evaluation information;

[0054] A selection module is used to select m target satisfaction evaluation attribute information with larger weights from the satisfaction evaluation attribute information, where m is a positive integer;

[0055] The acquisition module is also used to obtain the correlation accuracy of each target satisfaction evaluation attribute information;

[0056] A determination module, used to determine the target weight of each target satisfaction evaluation attribute information according to the weight and the associated accuracy of each target satisfaction evaluation attribute information and the preset relationship, wherein the preset relationship includes the relationship between the weight, the associated accuracy and the target weight of the preset satisfaction evaluation attribute information, and the preset satisfaction evaluation attribute information includes m target satisfaction evaluation attribute information;

[0057] The determination module is also used to determine a preset number of target satisfaction evaluation attribute information with a larger target weight as key satisfaction evaluation attribute information.

[0058] In a third aspect, an embodiment of the present application provides an electronic device, the device comprising:

[0059] a processor and a memory storing computer program instructions;

[0060] The method for determining the attribute information of any one of the above items is implemented when the processor executes the computer program instructions.

[0061] In a fourth aspect, an embodiment of the present application provides a computer storage medium, on which computer program instructions are stored. When the computer program instructions are executed by a processor, the method for determining attribute information of any one of the above items is implemented.

[0062] In a fifth aspect, an embodiment of the present application provides a computer program product, characterized in that when the instructions in the computer program product are executed by a processor of an electronic device, the electronic device is enabled to execute any of the above-mentioned methods for determining attribute information.

[0063] The method, apparatus, device and computer storage medium for determining attribute information of the embodiments of the present application, the method comprising: obtaining multiple satisfaction evaluation information of a target time period and the weight of the satisfaction evaluation attribute information in each satisfaction evaluation information; selecting m target satisfaction evaluation attribute information with larger weights from the satisfaction evaluation attribute information, where m is a positive integer; obtaining the correlation accuracy of each target satisfaction evaluation attribute information; determining the target weight of each target satisfaction evaluation attribute information according to a preset relationship based on the weight and correlation accuracy of each target satisfaction evaluation attribute information, wherein the preset relationship comprises the relationship between the weight, correlation accuracy and target weight of the preset satisfaction evaluation attribute information, and the preset satisfaction evaluation attribute information comprises m target satisfaction evaluation attribute information; determining a preset number of target satisfaction evaluation attribute information with larger target weights as key satisfaction evaluation attribute information.

[0064] In this way, from the satisfaction evaluation attribute information in the multiple satisfaction evaluation information in the target time period, the target satisfaction evaluation attribute information with a larger weight is selected, and the target weight is determined based on the target satisfaction evaluation attribute information, the weight and the associated accuracy, and the target satisfaction evaluation attribute information with a larger target weight is selected as the key satisfaction evaluation attribute information. The key satisfaction evaluation attribute information is selected by comprehensive weight and associated accuracy, and no longer relies on manual methods, which not only saves labor costs, but also improves the accuracy of determining the key satisfaction evaluation attribute information. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] In order to more clearly illustrate the technical solution of the embodiments of the present application, the following is a brief introduction to the drawings required for use in the embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0066] Figure 1 is a flowchart of a method for determining attribute information provided by an embodiment of the present application;

[0067] Figure 2 is a flowchart of a method for determining attribute information provided by another embodiment of the present application;

[0068] Figure 3 is a flowchart of a method for determining attribute information provided by another embodiment of the present application;

[0069] Figure 4 is a flowchart of a method for determining attribute information provided by yet another embodiment of the present application;

[0070] Figure 5 is a flowchart of a method for determining attribute information provided by yet another embodiment of the present application;

[0071] Figure 6 is a flowchart of a method for determining attribute information provided by yet another embodiment of the present application;

[0072] Figure 7 is a flowchart of a method for determining attribute information provided by yet another embodiment of the present application;

[0073] Figure 8 is a structural diagram of a system for determining attribute information provided by yet another embodiment of the present application;

[0074] Fig. 9 is a schematic diagram of the structure of a device for determining attribute information provided in yet another embodiment of the present application;

[0075] Fig.10 It is a structural diagram of an electronic device provided in yet another embodiment of the present application. DETAILED DESCRIPTION

[0076] The features and exemplary embodiments of various aspects of the present application will be described in detail below. In order to make the purpose, technical solutions and advantages of the present application clearer, the present application will be further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only intended to explain the present application, rather than to limit the present application. For those skilled in the art, the present application can be implemented without the need for some of these specific details. The following description of the embodiments is only to provide a better understanding of the present application by illustrating the examples of the present application.

[0077] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the statement "include..." do not exclude the presence of other identical elements in the process, method, article or device including the elements.

[0078] Operators need to consider user satisfaction while paying attention to the growth of the number of users. If user growth is based on sacrificing user usage perception, giving users a poor service experience is not conducive to the long-term development of the business. Therefore, the existing service operation system pays more and more attention to improving user satisfaction. In order to improve user satisfaction, it is necessary to analyze user satisfaction evaluation information.

[0079] In the prior art, the acquired satisfaction evaluation information is usually studied manually, and key satisfaction evaluation attribute information is extracted from the satisfaction evaluation information to determine the factors affecting user satisfaction. This method of determining key satisfaction evaluation attribute information not only has a relatively high labor cost, but also has a large degree of subjectivity, and the key satisfaction evaluation attribute information is not accurately determined.

[0080] In order to solve the problems of the prior art, the embodiments of the present application provide a method, device, equipment and computer storage medium for determining attribute information. The method for determining attribute information provided by the embodiments of the present application selects target satisfaction evaluation attribute information with a larger weight from the satisfaction evaluation attribute information in multiple satisfaction evaluation information in the target time period, and determines the target weight based on the target satisfaction evaluation attribute information, the weight and the associated accuracy, and selects the target satisfaction evaluation attribute information with a larger target weight as the key satisfaction evaluation attribute information. The key satisfaction evaluation attribute information is selected by comprehensive weight and associated accuracy, and no longer relies on manual methods, which not only saves labor costs, but also improves the accuracy of determining the key satisfaction evaluation attribute information.

[0081] The following first introduces the method for determining attribute information provided in the embodiment of the present application. Figure 1 A flowchart of a method for determining attribute information provided by an embodiment of the present application is shown.

[0082] like Figure 1As shown, the method for determining attribute information provided in the embodiment of the present application includes the following steps.

[0083] S110. Obtain multiple satisfaction evaluation information of a target time period and the weight of the satisfaction evaluation attribute information in each satisfaction evaluation information.

[0084] Here, the target time period is variable. The target time period is set in advance. The weight of the satisfaction evaluation attribute information can be set in advance.

[0085] In some embodiments, at each target time period, online user satisfaction evaluation information within the previous target time period is collected.

[0086] In some embodiments, the length of the satisfaction evaluation information is limited, and the satisfaction evaluation information that is too long or too short in the target time period is filtered out, and the satisfaction evaluation information containing special characters is filtered out. Specifically, the satisfaction evaluation information whose length is less than the first preset value or whose length is greater than the second preset value is filtered out. The special characters are set in advance.

[0087] In some embodiments, the weight of the satisfaction evaluation attribute information is a weight of at least one dimension, for example, a weight set in advance based on user experience, a weight related to information entropy, a weight related to importance, and a weight related to satisfaction level.

[0088] S120. Select m target satisfaction evaluation attribute information with larger weights from the satisfaction evaluation attribute information, where m is a positive integer.

[0089] In some embodiments, the satisfaction evaluation attribute information is sorted according to the size of the weight to obtain a satisfaction evaluation attribute information sequence. The first m target satisfaction evaluation attribute information with larger weights are selected from the satisfaction evaluation attribute information sequence.

[0090] In some embodiments, the value of m may be set in advance, that is, the amount of target satisfaction evaluation attribute information may be set in advance.

[0091] In some embodiments, when selecting target satisfaction evaluation attribute information, when the sum of the weights of the selected target satisfaction evaluation attribute information is greater than a third preset value, m target satisfaction evaluation attribute information is obtained.

[0092] S130, obtaining the correlation accuracy of each target satisfaction evaluation attribute information.

[0093] Here, the correlation accuracy of the target satisfaction evaluation attribute information refers to the accuracy of predicting the user satisfaction level based on the target satisfaction evaluation attribute information, and may be a value set in advance based on user experience.

[0094] S140. According to the weight and associated accuracy of each target satisfaction evaluation attribute information, the target weight of each target satisfaction evaluation attribute information is determined according to a preset relationship. The preset relationship includes the relationship between the weight, associated accuracy and target weight of the preset satisfaction evaluation attribute information. The preset satisfaction evaluation attribute information includes m target satisfaction evaluation attribute information.

[0095] Here, the weight of the target satisfaction evaluation attribute information is used as the secondary attribute weight, and the target weight of the target satisfaction evaluation attribute information is used as the primary attribute weight. The weight can represent the importance score of the satisfaction evaluation attribute information.

[0096] In some embodiments, for each target satisfaction evaluation attribute information, based on its associated accuracy, further adjustment is made according to the secondary attribute weight to obtain the final primary attribute weight. When the secondary attribute weight of each dimension of the target satisfaction evaluation attribute information is larger, the primary attribute weight should also be larger, and the weight of the target satisfaction evaluation attribute information should be increased; when the secondary attribute weights of different dimensions are very different, it proves that this attribute may not be stable enough, and the weight of the target satisfaction evaluation attribute information should be reduced.

[0097] As an example, when the weight of the target satisfaction evaluation attribute information is three-dimensional, the preset relationship is as shown in formula (1), which is as follows:

[0098]

[0099] Among them, U i , K i , Z i are the weights of the three dimensions of the target satisfaction evaluation attribute information; w1, w2, w3 correspond to the importance coefficients of the weights of the three dimensions, which are set in advance; v1, v2, v3 are used to control the degree of penalty when there are differences between different weights, which are set in advance; α is the smoothing factor; P i ' is the target weight; P i is the association accuracy. Weight enhancement factor Controls the positive correlation between the final primary attribute weight and the secondary attribute weight, v1(U i -K i ) 2 +v2(U i -Z i ) 2 +v3(U i -Z i ) 2 Control the final negative correlation between the primary attribute weight and the secondary attribute weight. When the difference between the secondary attribute weights of different dimensions is greater, v1(Ui -K i ) 2 +v2(U i -Z i ) 2 +v3(U i -Z i ) 2 The larger the value, the smaller the first-level attribute weight Pi' corresponding to the attribute.

[0100] S150: Determine a preset number of target satisfaction evaluation attribute information with a larger target weight as key satisfaction evaluation attribute information.

[0101] Here, the preset quantity is set in advance.

[0102] In some embodiments, after determining that a preset number of target satisfaction evaluation attribute information with a larger target weight is key satisfaction evaluation attribute information, the user's query input for the key satisfaction evaluation attribute information of the target time period is received, and the key satisfaction evaluation attribute information and its target weight are displayed for user analysis.

[0103] In this way, from the satisfaction evaluation attribute information in the multiple satisfaction evaluation information in the target time period, the target satisfaction evaluation attribute information with a larger weight is selected, and the target weight is determined based on the target satisfaction evaluation attribute information, the weight and the associated accuracy, and the target satisfaction evaluation attribute information with a larger target weight is selected as the key satisfaction evaluation attribute information. The key satisfaction evaluation attribute information is selected by comprehensive weight and associated accuracy, and no longer relies on manual methods, which not only saves labor costs, but also improves the accuracy of determining the key satisfaction evaluation attribute information.

[0104] Based on this, in some embodiments, the weight includes a first sub-weight, a second sub-weight and a third sub-weight; Figure 2 As shown, the above S120 may specifically include:

[0105] S121, from the satisfaction evaluation attribute information, select k first satisfaction evaluation attribute information with larger weights in the first sub-weights, l second satisfaction evaluation attribute information with larger weights in the second sub-weights, and n third satisfaction evaluation attribute information with larger weights in the third sub-weights;

[0106] S122, combining the first satisfaction evaluation attribute information, the second satisfaction evaluation attribute information and the third satisfaction evaluation attribute information to obtain m target satisfaction evaluation attribute information;

[0107] Among them, the sum of the first sub-weights of the k first satisfaction evaluation attribute information, the sum of the second sub-weights of the l second satisfaction evaluation attribute information, and the sum of the third sub-weights of the n third satisfaction evaluation attribute information are all not less than the first threshold; m≤k+l+n.

[0108] In some embodiments, for the weights of the three dimensions, namely the first sub-weight, the second sub-weight and the third sub-weight, the satisfaction evaluation attribute information with higher weights is selected respectively to generate an attribute set. For each dimension a, an attribute set is generated.

[0109] Specifically, the sub-weights of each satisfaction evaluation attribute information under dimension a are sorted to obtain Wd1={w d1 ,w d2 ,...,w dn}, where n is the number of satisfaction evaluation attribute information, w di It represents the sub-weight of the i-th satisfaction evaluation attribute information after sorting the sub-weights of the satisfaction evaluation attribute information corresponding to dimension a. The number of selected satisfaction evaluation attribute information k is determined by formula (2): d , formula (2) is as follows:

[0110]

[0111] Among them, T represents the first threshold, which is set in advance; the first k satisfaction evaluation attribute information is selected so that the sum of the weights of these k attributes is greater than T.

[0112] It can be understood that for the first sub-weight, k first satisfaction evaluation attribute information is selected; for the second sub-weight, l second satisfaction evaluation attribute information is selected; for the third sub-weight, n third satisfaction evaluation attribute information is selected. d The satisfaction evaluation attribute information and the first k d Sub-weights corresponding to the satisfaction evaluation attribute information Here, the number of satisfaction evaluation attribute information selected for each attribute a is k d are not necessarily the same. d Including k, l and n. The satisfaction evaluation attribute information selected in the three dimensions is combined to obtain the attribute set D merge ={d1,d1,...,d m} and weight set W merge ={w1,w2,...w m}, where m is the number of target satisfaction evaluation attribute information after the union of the satisfaction evaluation attribute information selected in the three dimensions.

[0113] In this way, the sub-weights of the three dimensions are comprehensively used to select the target satisfaction evaluation attribute information, which further improves the accuracy of the target satisfaction evaluation attribute information selection, and then improves the accuracy of the key satisfaction evaluation attribute information determination.

[0114] Based on this, in some embodiments, in the above S110, obtaining the weight of the satisfaction evaluation attribute information in each satisfaction evaluation information includes:

[0115] Obtain the frequency of the satisfaction evaluation attribute information in each satisfaction evaluation information, the evaluation level of the satisfaction evaluation attribute information in each satisfaction evaluation information, and the probability of the satisfaction evaluation attribute information in the satisfaction evaluation information of each satisfaction level;

[0116] According to the frequency of the satisfaction evaluation attribute information, determine the proportion of the satisfaction evaluation attribute information in each satisfaction evaluation information;

[0117] According to the proportion of the satisfaction evaluation attribute information in each satisfaction evaluation information, the entropy of the satisfaction evaluation attribute information is determined;

[0118] Determine a first sub-weight of the satisfaction evaluation attribute information according to the entropy of the satisfaction evaluation attribute information;

[0119] determining an importance index of the satisfaction evaluation attribute information according to the evaluation level of the satisfaction evaluation attribute information in each satisfaction evaluation information and the number of the plurality of satisfaction evaluation attribute information;

[0120] Normalizing the importance index of the satisfaction evaluation attribute information to obtain the second sub-weight of the satisfaction evaluation attribute information;

[0121] The third sub-weight of the satisfaction evaluation attribute information is determined according to the probability of the satisfaction evaluation attribute information in the satisfaction evaluation information of each satisfaction level.

[0122] In some embodiments, corresponding to the first sub-weight, the frequency of the satisfaction evaluation attribute information in each satisfaction evaluation information is obtained to construct an attribute matrix of the online evaluation data:

[0123]

[0124] Among them, g is the number of selected attributes, s is the number of evaluation data, and x sg It represents the frequency of the gth satisfaction evaluation attribute information in the sth satisfaction evaluation information. It should be noted that the frequency is calculated for each satisfaction evaluation information. Formula (3) is used to determine the proportion of each satisfaction evaluation attribute information in each satisfaction evaluation information. Formula (3) is as follows:

[0125]

[0126] Formula (3) can be understood as standardizing the frequency of the satisfaction evaluation attribute information in each satisfaction evaluation information to ensure that the sum of all frequencies in each column is 1.

[0127] According to the proportion of satisfaction evaluation attribute information in each satisfaction evaluation information, the entropy of satisfaction evaluation attribute information is determined using formula (4), Q j represents the entropy of the jth satisfaction evaluation attribute information, and formula (4) is as follows:

[0128]

[0129] Formula (5) is used to calculate the entropy redundancy of each satisfaction evaluation attribute information, W j represents the entropy redundancy of the jth attribute, and formula (5) is as follows:

[0130] W j =1-Q j (5)

[0131] Among them, entropy Q j The larger the entropy redundancy W is, the j The smaller it is, the weaker the distinguishing ability of the satisfaction evaluation attribute information is. Conversely, the stronger the distinguishing ability of the satisfaction evaluation attribute information is, the more important the satisfaction evaluation attribute information is.

[0132] According to the entropy of the satisfaction evaluation attribute information, the first sub-weight of the satisfaction evaluation attribute information is determined using formula (6). j represents the first sub-weight of the j-th attribute, and formula (6) is as follows:

[0133]

[0134] In some embodiments, corresponding to the second sub-weight, the evaluation level of the satisfaction evaluation attribute information in each satisfaction evaluation information is obtained, and the importance index of the satisfaction evaluation attribute information is determined according to the evaluation level of the satisfaction evaluation attribute information in each satisfaction evaluation information and the number of multiple satisfaction evaluation attribute information. For the satisfaction evaluation attribute information j, its importance index is calculated using formula (7), and formula (7) is as follows:

[0135]

[0136] Among them, E ijis the evaluation level of the i-th satisfaction evaluation attribute information in the satisfaction evaluation information of the i-th user. This evaluation level is determined based on the user's subjective experience and can be directly obtained. It is the value entered by the user. In this part, the importance of the satisfaction evaluation attribute information is scored by multiple evaluators. d is the number of evaluators, and g is the number of satisfaction evaluation attribute information.

[0137] The importance index E of the satisfaction evaluation attribute information is calculated using formula (8). j Normalized to obtain the second sub-weight K of each satisfaction evaluation attribute information j j , where g is the number of satisfaction evaluation attribute information. Formula (8) is as follows:

[0138]

[0139] In some embodiments, corresponding to the third sub-weight, the probability of the satisfaction evaluation attribute information in the satisfaction evaluation information of each satisfaction level is obtained. Here, the satisfaction level of the satisfaction evaluation information is known, which can be a label directly obtained manually, or can be predicted by the method provided in the embodiment of the present application. For multiple satisfaction evaluation information, the probability of each satisfaction evaluation attribute information j appearing in each satisfaction level y is counted, and the statistical formula (9) is as follows:

[0140]

[0141] Among them, t y,j It indicates the number of times the satisfaction evaluation attribute information j appears when the satisfaction level is y, and Y indicates the number of satisfaction levels.

[0142] For each satisfaction evaluation attribute information j, the Gini coefficient G is calculated using formula (10): j , formula (10) is as follows:

[0143]

[0144] The Gini coefficient of the satisfaction evaluation attribute information is normalized using formula (11) to obtain the third sub-weight of each satisfaction evaluation attribute information. Formula (11) is as follows:

[0145]

[0146] Among them, g is the number of satisfaction evaluation attribute information.

[0147] In this way, considering different dimensions, the weights of the satisfaction evaluation attribute information are determined respectively, including the weight of the satisfaction evaluation attribute information calculated based on information entropy, the weight of the satisfaction evaluation attribute information calculated based on domain knowledge and experience, and the weight of the satisfaction evaluation attribute information calculated based on the Gini coefficient. This provides a basis for the calculation of the secondary attribute weights and improves the accuracy of determining the target weights.

[0148] Based on this, in some embodiments, Figure 3 As shown, before the above S130, the method may further include:

[0149] S123, randomly selecting multiple first target satisfaction evaluation attribute information from the m target satisfaction evaluation attribute information;

[0150] S124, for each selection of multiple first target satisfaction evaluation attribute information, respectively use the multiple first target satisfaction evaluation attribute information to construct a first satisfaction prediction model, and input the multiple satisfaction evaluation information into the first satisfaction prediction model, and use the first satisfaction classification rules corresponding to the multiple first target satisfaction evaluation attribute information in the first satisfaction prediction model to predict the satisfaction levels corresponding to the multiple satisfaction evaluation information, so as to obtain the predicted satisfaction levels corresponding to the multiple satisfaction evaluation information;

[0151] S125, determining the accuracy of the first satisfaction prediction model according to the predicted satisfaction levels and the actual satisfaction levels corresponding to the plurality of satisfaction evaluation information;

[0152] S126, calculating the average value of the accuracy of the first satisfaction prediction model to obtain an average accuracy;

[0153] S127. Determine the associated accuracy of each target satisfaction evaluation attribute information according to the average accuracy and the amount of the first target satisfaction evaluation attribute information.

[0154] Specifically, for D merge For each target satisfaction evaluation attribute information in , calculate the correlation accuracy P between r (1 <= r <= m) satisfaction evaluation attribute information i,r , the specific process is as follows:

[0155] Step 1. From D merge Randomly select r satisfaction evaluation attribute information (including satisfaction evaluation attribute information d i ).

[0156] Step 2. Based on the selected r satisfaction evaluation attribute information, use multiple first target satisfaction evaluation attribute information to construct a first satisfaction prediction model, and input multiple satisfaction evaluation information into the first satisfaction prediction model. Use the first satisfaction classification rules corresponding to multiple first target satisfaction evaluation attribute information in the first satisfaction prediction model to predict the satisfaction levels corresponding to the multiple satisfaction evaluation information, and obtain the predicted satisfaction levels corresponding to the multiple satisfaction evaluation information. Determine the accuracy of the first satisfaction prediction model based on the predicted satisfaction levels and actual satisfaction levels corresponding to the multiple satisfaction evaluation information. Here, the accuracy can also be replaced by model effect indicators such as recall, F1, and precision.

[0157] Step 3. Repeat the above steps 1 and 2 t times to obtain the accuracy of t first satisfaction prediction models: {F r,1 ,F r,2 ,...,F r,t}.

[0158] Step 4. Use formula (12) to calculate the average accuracy of the t first satisfaction prediction models as the correlation accuracy P of the r satisfaction evaluation attribute information. i,r , formula (12) is as follows:

[0159]

[0160] Afterwards, for D merge Each satisfaction evaluation attribute information d i , use formula (13) to calculate the correlation accuracy P between the satisfaction evaluation attribute information and all other satisfaction evaluation attribute information i , formula (13) is as follows:

[0161]

[0162] Among them, P i,r P represents the correlation accuracy between r satisfaction evaluation attribute information. i,r From the perspective of the combination of multiple satisfaction evaluation attribute information, the importance of a satisfaction evaluation attribute information is measured. Compared with the three secondary attribute weights, the effect of each satisfaction evaluation attribute information is measured more comprehensively.

[0163] It should be noted that when multiple random selections are made, the value of r changes, and can be from 1 to m.

[0164] In this way, the effect of each satisfaction evaluation attribute information is measured more comprehensively through the correlation accuracy between the satisfaction evaluation attribute information, thereby improving the accuracy of the key satisfaction evaluation attribute information.

[0165] Based on this, in some embodiments, Figure 4 As shown, before obtaining the weight of the satisfaction evaluation attribute information in each satisfaction evaluation information, the method may further include:

[0166] S101, segmenting multiple satisfaction evaluation information to obtain multiple target segmentations;

[0167] S102, deleting the segmentations that are consistent with the preset stop words from the multiple target segmentations to obtain multiple first segmentations;

[0168] S103: Delete the first participles whose appearance frequency in the plurality of satisfaction evaluation information is less than the second threshold, to obtain satisfaction evaluation attribute information.

[0169] Here, the preset stop words are set in advance, indicating that the words cannot be used as attribute information for satisfaction evaluation. The second threshold is set in advance.

[0170] In this way, stop words and less frequently occurring words are removed, the amount of calculation is reduced, and computing resources are saved.

[0171] Based on this, in some embodiments, Figure 5 As shown, the above S103 may specifically include:

[0172] S104, deleting the first participles whose frequency of appearance in the plurality of satisfaction evaluation information is less than a second threshold, to obtain a plurality of second participles;

[0173] S105, for each target second participle in the plurality of second participles, calculating the similarity between the target second participle and a third participle, where the third participle is a participle in the plurality of second participles except the target second participle;

[0174] S106. Delete, from the plurality of second participles, third participles whose similarity to the target second participle is greater than a third threshold, to obtain satisfaction evaluation attribute information.

[0175] Here, the third threshold is set in advance.

[0176] In some embodiments, multiple second word segments are vectorized to obtain E={e1, e2, ..., e x}, the vectorization method used is not limited, and can include but is not limited to Word2Vec. Sort the second participles by frequency from large to small; for the i-th participle in the sorted second participles, calculate the cosine similarity cos between the j-th participle and the i-th participle ij , when cos ij If it is less than a certain value, the second word j is filtered. The calculation formula (14) of cosine similarity is as follows:

[0177]

[0178] Among them, e i,c represents the value of the cth dimension of the i-th second word segmentation vector, e j,c represents the value of the cth dimension of the jth second word vector, and h represents the dimension of the word vector.

[0179] It should be noted that although this step deletes the third participles whose similarity to the target second participle is greater than the third threshold, thus reducing the number of participles, it is necessary to save the mapping relationship between the deleted participles and the participles similar to them. i,j If it is greater than the third threshold, (e j ,e i ) is saved in the mapping table M. The second word segmentation e is processed in the subsequent stage j When it is used, it is regarded as the second participle e i to handle.

[0180] In this way, for segmented words with high similarity, the repeated parts are deleted to avoid the appearance of synonyms and near-synonyms, and only a representative segmented word is retained as the attribute information of satisfaction evaluation, which further reduces the amount of calculation and saves resources.

[0181] Based on this, in some embodiments, Figure 6 As shown, the above S106 may specifically include:

[0182] S107, deleting, from the plurality of second participles, third participles whose similarity to the target second participle is greater than a third threshold, to obtain a plurality of fourth participles;

[0183] S108: Select a fourth word whose appearance frequency in a plurality of preset words is greater than a fourth threshold value to obtain satisfaction evaluation attribute information.

[0184] Here, the preset vocabulary is set in advance and may be some data that is frequently used in daily life.

[0185] Specifically, set the number of topics {t1, t2, ..., t q}, maximum number of iterations I. For each topic number tq, train an LDA topic model. Through this topic model, we can get each fourth participle e i Belong to any topic j The probability p(e i / topic j ), where any topic topicj includes multiple preset words. Further, for each fourth participle e i , and get its topic vector representation:

[0186] P i =[p(ei / topic1), p(e i / topic2), ..., p(e i / topic q*I )].

[0187] Perform k-means clustering on the topic vectors of all fourth participles to obtain the fourth participle set of each category. Combine the vocabulary clustering results under different topics, select the fourth participles whose frequency of occurrence in different topics is greater than the fourth threshold, and obtain the satisfaction evaluation attribute information.

[0188] In some embodiments, an artificial condensation method may also be used to determine the meaning of multiple preset words based on a topic, and select appropriate word segments as satisfaction evaluation attribute information used in satisfaction analysis.

[0189] In this way, multiple preset words are used as some data that are used more frequently in daily life. Through the preset words, the frequently appearing words are used as the attribute information of satisfaction evaluation, the noise is removed, and the words that are strongly related to satisfaction are selected, which further reduces the amount of calculation and saves resources.

[0190] Based on this, in some embodiments, Figure 7 As shown, the method may also include:

[0191] S210, comparing each satisfaction evaluation information with the satisfaction evaluation attribute information to obtain a comparison result;

[0192] S220, generating characteristic information of each satisfaction evaluation information according to the comparison result;

[0193] S230. Input the characteristic information into the second satisfaction prediction model, and predict the satisfaction levels corresponding to the multiple satisfaction evaluation information according to the second satisfaction classification rule in the second satisfaction prediction model to obtain the satisfaction levels corresponding to the multiple satisfaction evaluation information.

[0194] Here, the comparison result may include attribute information included in the satisfaction evaluation information that is consistent with the satisfaction evaluation attribute information, or may include attribute information not included in the satisfaction evaluation information that is consistent with the satisfaction evaluation attribute information.

[0195] In some embodiments, after the satisfaction evaluation information is preprocessed and the satisfaction evaluation attribute information is screened, the satisfaction level of the processed satisfaction evaluation information is predicted using the CART model. The CART model is a very effective non-parametric classification and regression method, first proposed by Breiman et al., and has been widely used in the field of statistics and data mining technology. The advantage of CART is that it is easy to understand, use and interpret.

[0196] Specifically, the satisfaction evaluation information to be predicted is first segmented, and then features are extracted based on the segmented results. For the satisfaction evaluation attribute information, an experimental attribute set is constructed. For each word in the experimental attribute set, if it appears in the satisfaction evaluation information, the feature of the satisfaction evaluation information corresponding to the satisfaction evaluation attribute information is 1, otherwise it is 0. The feature values ​​of each satisfaction evaluation attribute information are concatenated together to obtain feature information. For example, if the feature information of a satisfaction evaluation attribute information is [1,0,1,...,1], it means that the first satisfaction evaluation attribute information, the third satisfaction evaluation attribute information and the last satisfaction evaluation attribute information appear in the current satisfaction evaluation information, and the second satisfaction evaluation attribute information does not appear in the current satisfaction evaluation information.

[0197] It should be noted that the order of the satisfaction evaluation attribute information in the experimental attribute set is fixed.

[0198] In some embodiments, after feature extraction is performed on the satisfaction evaluation information to be predicted, the extracted feature information is input into a second satisfaction prediction model (eg, a CART tree) to predict the satisfaction level of the satisfaction evaluation information.

[0199] In this way, based on the satisfaction evaluation attribute information in the satisfaction evaluation information, the satisfaction level of the satisfaction evaluation information can be predicted. By predicting the satisfaction level of the satisfaction evaluation information of a batch of users, the overall satisfaction distribution of users can be known, and operators can understand their own service quality.

[0200] Based on this, in some embodiments, before the above S230, the method may further include:

[0201] Obtaining multiple satisfaction evaluation information samples and actual satisfaction levels corresponding to the satisfaction evaluation information samples;

[0202] Segmenting multiple satisfaction evaluation information samples to obtain multiple satisfaction evaluation attribute information samples;

[0203] Compare each satisfaction evaluation information sample with multiple satisfaction evaluation attribute information samples to obtain a sample comparison result;

[0204] According to the sample comparison results, a characteristic information sample of each satisfaction evaluation information is generated;

[0205] Inputting the characteristic information samples into the initial second satisfaction prediction model, predicting the satisfaction levels corresponding to the multiple satisfaction evaluation information samples according to the initial second satisfaction classification rules in the initial second satisfaction prediction model, and obtaining the sample predicted satisfaction levels corresponding to the multiple satisfaction evaluation information samples;

[0206] Determining the accuracy of the initial second satisfaction prediction model according to the sample prediction satisfaction levels corresponding to the multiple satisfaction evaluation information samples and the actual satisfaction levels of the multiple satisfaction evaluation information samples;

[0207] Update the initial second satisfaction classification rules in the initial second satisfaction prediction model, predict the satisfaction levels corresponding to multiple satisfaction evaluation information samples according to the updated initial second satisfaction classification rules, and obtain updated sample predicted satisfaction levels; update the accuracy of the initial second satisfaction prediction model according to the updated sample predicted satisfaction levels and the actual satisfaction levels of the multiple satisfaction evaluation information samples; when the accuracy of the initial second satisfaction prediction model is greater than the fourth threshold, obtain the second satisfaction prediction model, and the second satisfaction prediction model includes the second satisfaction classification rules.

[0208] Here, the fourth threshold is set in advance.

[0209] In some embodiments, the CART tree is trained according to the Gini coefficient. The Gini coefficient is used to determine the best splitting node of the tree. In the process of constructing the CART tree, by continuously splitting nodes, the Gini coefficient of each child node is made as small as possible, thereby enabling different categories of satisfaction evaluation information samples to be better divided.

[0210] In this way, the initial second satisfaction prediction model is trained with a large number of samples. When the accuracy of the initial second satisfaction prediction model is greater than the fourth threshold, the obtained second satisfaction prediction model can more accurately predict the satisfaction level of the satisfaction evaluation information.

[0211] In the embodiments provided in this application, a machine learning method is used to evaluate user satisfaction, which improves the efficiency and accuracy of the evaluation and reduces the cost of the evaluation compared to the method based on manual evaluation; compared to the existing machine learning-based method, the key satisfaction evaluation attribute information that affects satisfaction is automatically selected by performing multi-level analysis on text data. Using these key satisfaction evaluation attribute information for evaluation and analysis can not only improve the accuracy and efficiency of satisfaction evaluation, but also associate satisfaction with related satisfaction evaluation attribute information, providing clues for improving satisfaction.

[0212] In the embodiment provided by the present application, by selecting the satisfaction evaluation attribute information based on statistical and semantic features, the calculation efficiency can be improved, while focusing on key words and improving the accuracy of the analysis. Through the multi-level and multi-dimensional satisfaction evaluation attribute information weight analysis, it can help operators find factors that affect the evaluation. For the satisfaction evaluation attribute information weight analysis stage, when calculating the primary attribute weight, not only multiple secondary attribute weights are considered, but also the relationship between multiple satisfaction evaluation attribute information is used to comprehensively evaluate each satisfaction evaluation attribute information, so that the weight evaluation of the satisfaction evaluation attribute information is more reasonable. When converting the correlation accuracy rate to the final primary attribute weight, on the basis of the correlation accuracy rate, the weight of the satisfaction evaluation attribute information is effectively adjusted through the weight enhancement factor and the weight reduction factor. Not only the contribution of the sub-dimensional weight to the final weight is considered, but also the impact on the final weight when the sub-dimensional weight difference is large is considered. The analysis and feature mining of user satisfaction by operators is realized, the efficiency of data information utilization is improved, the efficiency and ability of satisfaction analysis and evaluation are improved, and the full utilization of massive, multi-dimensional, incomplete actual data can be ensured, and the ability to mine user satisfaction-related features is also enhanced.

[0213] Based on the above embodiments, the embodiments of the present application are explained in detail below, for example:

[0214] (1) The original input satisfaction evaluation information is as follows:

[0215] Satisfaction evaluation information 1: "The new package launched in January 2023 is too expensive and the Internet speed has become much slower. It is not as good as last year's package, which has affordable pricing and fast Internet speed."

[0216] Satisfaction evaluation information 2: "I like the March 2023 package, it is cost-effective"

[0217] Satisfaction evaluation information 3: "Hahaha, la la la, ****, 888"

[0218] Satisfaction evaluation information 4: “Hey hey”

[0219] Satisfaction evaluation information 5: “!@@@@@@@@@”

[0220] (2) Data preprocessing: Filter satisfaction evaluation information 3, satisfaction evaluation information 4, satisfaction evaluation information 5, and the remaining satisfaction evaluation information 1 and satisfaction evaluation information 2. Segment the satisfaction evaluation information 1 and satisfaction evaluation information 2, and obtain the following results:

[0221] Satisfaction evaluation attribute information of satisfaction evaluation information 1:

[0222] Word segmentation results: [2023, January, new, launched, package, price, too expensive, internet speed, slowed down, a lot, not as good as last year, package, pricing, affordable, internet speed, fast]

[0223] Actual satisfaction level: 0

[0224] Satisfaction evaluation attribute information of satisfaction evaluation information 2:

[0225] Word segmentation results: [I, like, 2023, March, package, cost-effective]

[0226] Actual satisfaction level: 1

[0227] (3) Fusion of multiple satisfaction evaluation attribute information: Based on satisfaction evaluation attribute information 1 and 2, multiple satisfaction evaluation attribute information is screened out:

[0228] [2023, January, launch, package, price, too expensive, internet speed, slow down, not as good as, affordable, fast, I, like, March, cost-effective]

[0229] (4) Satisfaction evaluation: The above-mentioned satisfaction evaluation information 1 and satisfaction evaluation information 2 are satisfaction evaluation information for predicting the satisfaction level. Based on the word segmentation results of the satisfaction evaluation information and multiple satisfaction evaluation attribute information, through feature extraction and prediction of the second satisfaction prediction model, the output results are as follows:

[0230] Satisfaction label prediction value of satisfaction evaluation information 1: 0

[0231] Satisfaction label prediction value of satisfaction evaluation information 2: 1

[0232] (5) Multi-level and multi-dimensional attribute information weight analysis to determine key satisfaction evaluation attribute information: Analyze multiple satisfaction evaluation attribute information that affect the satisfaction level to obtain key satisfaction evaluation information. Specifically, for the multiple satisfaction evaluation attribute information obtained in step (3), first obtain the secondary attribute weight of each satisfaction evaluation attribute information, and then calculate the primary attribute weight. The obtained secondary attribute weight includes a 3*15 secondary attribute weight matrix, where the i-th row and j-th column of the matrix is ​​a floating point number, indicating the weight of the i-th dimension corresponding to the j-th attribute. According to the above secondary attribute weight matrix, multiple satisfaction evaluation attribute information is screened to select more accurate target satisfaction evaluation attribute information, and obtain the attribute set: [price, too expensive, network, slow, not as good as, affordable, fast, like, cost-effective], and finally obtain a vector with a dimension of 10. The value of the vector is a floating point number, and the jth element of the vector represents the primary attribute weight of the corresponding jth attribute, such as [0.7, 0.8, 0.5, 0.9, 0.3, 0.7, 0.5, 0.6, 0.65, 0.95]. According to the vector, it can be known that the top three key satisfaction evaluation information that most affect the evaluation are:

[0233] The last satisfaction evaluation information: "High cost performance", target weight 0.95;

[0234] The fourth satisfaction rating message is “slowing down”, with a target weight of 0.9;

[0235] The second satisfaction evaluation information is “too expensive” and has a target weight of 0.8.

[0236] Furthermore, when operators want to improve customer satisfaction, they know that they need to improve the cost-effectiveness of services, especially increase network speed, and the price cannot be too high.

[0237] The method for determining attribute information provided in this application can be applied to a system for determining attribute information. Figure 8 As shown, the attribute information determination system 300 includes a data preprocessing module 310, an attribute selection module 320 based on statistical and semantic features, a satisfaction evaluation module 330 based on the CART algorithm, and a multi-level and multi-dimensional attribute weight analysis module 340. The data preprocessing module 310 is used to obtain satisfaction evaluation information, including filtering out satisfaction evaluation information that is too long or too short. The attribute selection module 320 based on statistical and semantic features obtains satisfaction evaluation attribute information in the satisfaction evaluation information. The satisfaction evaluation module 330 based on the CART algorithm is used to predict the satisfaction levels corresponding to multiple satisfaction evaluation information. The multi-level and multi-dimensional attribute weight analysis module 340 is used to determine key satisfaction evaluation attribute information.

[0238] Based on the method for determining attribute information provided in the above embodiment, the present application also provides a specific implementation of a device for determining attribute information. Please refer to the following embodiment.

[0239] See first Fig. 9 , the attribute information determination device 400 provided in the embodiment of the present application includes:

[0240] An acquisition module 410 is used to acquire multiple satisfaction evaluation information of a target time period and the weight of the satisfaction evaluation attribute information in each satisfaction evaluation information;

[0241] A selection module 420 is used to select m target satisfaction evaluation attribute information with larger weights from the satisfaction evaluation attribute information, where m is a positive integer;

[0242] The acquisition module 410 is also used to obtain the correlation accuracy of each target satisfaction evaluation attribute information;

[0243] A determination module 430 is used to determine the target weight of each target satisfaction evaluation attribute information according to the weight and the associated accuracy of each target satisfaction evaluation attribute information and the preset relationship, wherein the preset relationship includes the relationship between the weight, the associated accuracy and the target weight of the preset satisfaction evaluation attribute information, and the preset satisfaction evaluation attribute information includes m target satisfaction evaluation attribute information;

[0244] The determination module 430 is further used to determine a preset number of target satisfaction evaluation attribute information with a larger target weight as key satisfaction evaluation attribute information.

[0245] Based on this, in some embodiments, the weight includes a first sub-weight, a second sub-weight and a third sub-weight; the selection module 420 can be specifically used to:

[0246] From the satisfaction evaluation attribute information, select k first satisfaction evaluation attribute information with larger weights in the first sub-weight, l second satisfaction evaluation attribute information with larger weights in the second sub-weight, and n third satisfaction evaluation attribute information with larger weights in the third sub-weight;

[0247] Merging the first satisfaction evaluation attribute information, the second satisfaction evaluation attribute information and the third satisfaction evaluation attribute information to obtain m target satisfaction evaluation attribute information;

[0248] Among them, the sum of the first sub-weights of the k first satisfaction evaluation attribute information, the sum of the second sub-weights of the l second satisfaction evaluation attribute information, and the sum of the third sub-weights of the n third satisfaction evaluation attribute information are all not less than the first threshold; m≤k+l+n.

[0249] Based on this, in some embodiments, the acquisition module 410 may be specifically used to:

[0250] Obtain the frequency of the satisfaction evaluation attribute information in each satisfaction evaluation information, the evaluation level of the satisfaction evaluation attribute information in each satisfaction evaluation information, and the probability of the satisfaction evaluation attribute information in the satisfaction evaluation information of each satisfaction level;

[0251] According to the frequency of the satisfaction evaluation attribute information, determine the proportion of the satisfaction evaluation attribute information in each satisfaction evaluation information;

[0252] According to the proportion of the satisfaction evaluation attribute information in each satisfaction evaluation information, the entropy of the satisfaction evaluation attribute information is determined;

[0253] Determine a first sub-weight of the satisfaction evaluation attribute information according to the entropy of the satisfaction evaluation attribute information;

[0254] determining an importance index of the satisfaction evaluation attribute information according to the evaluation level of the satisfaction evaluation attribute information in each satisfaction evaluation information and the number of the plurality of satisfaction evaluation attribute information;

[0255] Normalizing the importance index of the satisfaction evaluation attribute information to obtain the second sub-weight of the satisfaction evaluation attribute information;

[0256] The third sub-weight of the satisfaction evaluation attribute information is determined according to the probability of the satisfaction evaluation attribute information in the satisfaction evaluation information of each satisfaction level.

[0257] Based on this, in some embodiments, the apparatus 400 may further include:

[0258] The selection module 420 is further configured to randomly select a plurality of first target satisfaction evaluation attribute information from the target satisfaction evaluation attribute information before obtaining the correlation accuracy of the target satisfaction evaluation attribute information;

[0259] The determination module 430 is further configured to construct a first satisfaction prediction model using the plurality of first target satisfaction evaluation attribute information respectively for each selection of the plurality of first target satisfaction evaluation attribute information, input the plurality of satisfaction evaluation information into the first satisfaction prediction model, and predict the satisfaction levels corresponding to the plurality of satisfaction evaluation information using the first satisfaction classification rules corresponding to the plurality of first target satisfaction evaluation attribute information in the first satisfaction prediction model, so as to obtain predicted satisfaction levels corresponding to the plurality of satisfaction evaluation information;

[0260] The determination module 430 is further used to determine the accuracy of the first satisfaction prediction model according to the predicted satisfaction levels and the actual satisfaction levels corresponding to the plurality of satisfaction evaluation information;

[0261] A calculation module, used to calculate the average accuracy of the first satisfaction prediction model to obtain an average accuracy;

[0262] The determination module 430 is further used to determine the associated accuracy of each target satisfaction evaluation attribute information according to the average accuracy and the amount of the first target satisfaction evaluation attribute information.

[0263] Based on this, in some embodiments, the apparatus 400 may further include:

[0264] A word segmentation module is used to segment multiple satisfaction evaluation information to obtain multiple target word segments before obtaining the weight of the satisfaction evaluation attribute information in each satisfaction evaluation information;

[0265] A deletion module is used to delete the segmentations that are consistent with the preset stop words from the multiple target segmentations to obtain multiple first segmentations;

[0266] The deletion module is also used to delete the first participles whose frequency of appearance in multiple satisfaction evaluation information is less than the second threshold, so as to obtain the satisfaction evaluation attribute information.

[0267] Based on this, in some embodiments, the deletion module can be specifically used to:

[0268] Deleting the first participles whose occurrence frequency in the plurality of satisfaction evaluation information is less than the second threshold, to obtain a plurality of second participles;

[0269] For each target second participle in the plurality of second participles, calculating the similarity between the target second participle and a third participle, where the third participle is a participle in the plurality of second participles except the target second participle;

[0270] From the plurality of second participles, third participles whose similarity to the target second participle is greater than a third threshold are deleted to obtain satisfaction evaluation attribute information.

[0271] Based on this, in some embodiments, the deletion module can be specifically used to:

[0272] From the plurality of second participles, delete the third participles whose similarity to the target second participle is greater than a third threshold, to obtain a plurality of fourth participles;

[0273] A fourth participle whose appearance frequency in a plurality of preset words is greater than a fourth threshold is selected to obtain satisfaction evaluation attribute information.

[0274] Based on this, in some embodiments, the apparatus 400 may further include:

[0275] A comparison module, used for comparing each satisfaction evaluation information with the satisfaction evaluation attribute information to obtain a comparison result;

[0276] A generation module, used for generating characteristic information of each satisfaction evaluation information according to the comparison result;

[0277] The determination module 430 is also used to input the characteristic information into the second satisfaction prediction model, predict the satisfaction levels corresponding to the multiple satisfaction evaluation information according to the second satisfaction classification rule in the second satisfaction prediction model, and obtain the satisfaction levels corresponding to the multiple satisfaction evaluation information.

[0278] Based on this, in some embodiments, the apparatus 400 may further include:

[0279] The acquisition module 410 is further used to acquire a plurality of satisfaction evaluation information samples and actual satisfaction levels corresponding to the satisfaction evaluation information samples before inputting the characteristic information into the second satisfaction prediction model;

[0280] The word segmentation module is also used to segment multiple satisfaction evaluation information samples to obtain multiple satisfaction evaluation attribute information samples;

[0281] The comparison module is also used to compare each satisfaction evaluation information sample with multiple satisfaction evaluation attribute information samples to obtain a sample comparison result;

[0282] The generation module is also used to generate characteristic information samples of each satisfaction evaluation information according to the sample comparison results;

[0283] The determination module 430 is further used to input the characteristic information sample into the initial second satisfaction prediction model, predict the satisfaction levels corresponding to the multiple satisfaction evaluation information samples according to the initial second satisfaction classification rule in the initial second satisfaction prediction model, and obtain the sample predicted satisfaction levels corresponding to the multiple satisfaction evaluation information samples;

[0284] The determination module 430 is further used to determine the accuracy of the initial second satisfaction prediction model according to the sample prediction satisfaction levels corresponding to the multiple satisfaction evaluation information samples and the actual satisfaction levels of the multiple satisfaction evaluation information samples;

[0285] Determination module 430 is also used to update the initial second satisfaction classification rules in the initial second satisfaction prediction model, predict the satisfaction levels corresponding to multiple satisfaction evaluation information samples according to the updated initial second satisfaction classification rules, and obtain updated sample predicted satisfaction levels; based on the updated sample predicted satisfaction levels and the actual satisfaction levels of multiple satisfaction evaluation information samples, update the accuracy of the initial second satisfaction prediction model; when the accuracy of the initial second satisfaction prediction model is greater than the fourth threshold, obtain the second satisfaction prediction model, and the second satisfaction prediction model includes the second satisfaction classification rules.

[0286] The various modules of the device for determining attribute information provided in the embodiment of the present application can implement the functions of the various steps of the method for determining attribute information provided above and achieve the corresponding technical effects, which will not be described in detail here for the sake of brevity.

[0287] Based on the same inventive concept, an embodiment of the present application also provides an electronic device.

[0288] Fig.10 A schematic diagram of the hardware structure of an electronic device provided in an embodiment of the present application is shown.

[0289] The electronic device may include a processor 501 and a memory 502 storing computer program instructions.

[0290] Specifically, the processor 501 may include a central processing unit (CPU), or an application specific integrated circuit (ASIC), or may be configured to implement one or more integrated circuits of the embodiments of the present application.

[0291] The memory 502 may include a large capacity memory for data or instructions. By way of example and not limitation, the memory 502 may include a hard disk drive (HDD), a floppy disk drive, a flash memory, an optical disk, a magneto-optical disk, a magnetic tape, or a universal serial bus (USB) drive or a combination of two or more of these. In appropriate cases, the memory 502 may include a removable or non-removable (or fixed) medium. In appropriate cases, the memory 502 may be inside or outside the integrated gateway disaster recovery device. In a specific embodiment, the memory 502 is a non-volatile solid-state memory.

[0292] The memory may include a read-only memory (ROM), a random access memory (RAM), a magnetic disk storage medium device, an optical storage medium device, a flash memory device, an electrical, optical or other physical / tangible memory storage device. Therefore, generally, the memory includes one or more tangible (non-transitory) computer-readable storage media (e.g., a memory device) encoded with software including computer-executable instructions, and when the software is executed (e.g., by one or more processors), it is operable to perform the operations described with reference to the method according to one aspect of the present disclosure.

[0293] The processor 501 implements any one of the attribute information determination methods in the above embodiments by reading and executing the computer program instructions stored in the memory 502 .

[0294] In one example, the electronic device may further include a communication interface 503 and a bus 510. Fig.10 As shown, the processor 501, the memory 502, and the communication interface 503 are connected via a bus 510 and communicate with each other.

[0295] The communication interface 503 is mainly used to implement communication between various modules, devices, units and / or equipment in the embodiments of the present application.

[0296] The bus 510 includes hardware, software or both, coupling the components of the electronic device to each other. For example, but not limitation, the bus may include an accelerated graphics port (Accelerated Graphics Port, AGP) or other graphics bus, an enhanced industry standard architecture (Extended Industry Standard Architecture, EISA) bus, a front side bus (Front Side Bus, FSB), a hypertransport (Hyper Transport, HT) interconnection, an industry standard architecture (Industry Standard Architecture, ISA) bus, an infinite bandwidth interconnection, a low pin count (Linear Predictive Coding, LPC) bus, a memory bus, a microchannel architecture (MicroChannel Architecture, MCA) bus, a peripheral component interconnect (Peripheral Component Interconnect, PCI) bus, a PCI-Express (Peripheral Component Interconnect-X, PCI-X) bus, a serial advanced technology attachment (Serial Advanced Technology Attachment, SATA) bus, a video electronics standard association local (VESA Local Bus, VLB) bus or other suitable buses or a combination of two or more of these. Where appropriate, the bus 510 may include one or more buses. Although the embodiments of the present application describe and illustrate a specific bus, the present application considers any suitable bus or interconnection. The electronic device can execute the method for determining attribute information in the embodiments of the present invention, thereby implementing the above-mentioned method for determining attribute information.

[0297] In addition, in combination with the method for determining attribute information in the above embodiments, the present application embodiment may provide a computer storage medium for implementation. The computer storage medium stores computer program instructions; when the computer program instructions are executed by a processor, any one of the methods for determining attribute information in the above embodiments is implemented.

[0298] The present application also provides a computer program product. When the instructions in the computer program product are executed by a processor of an electronic device, the electronic device executes each process of implementing any one of the above-mentioned attribute information determination method embodiments.

[0299] It should be clear that the present application is not limited to the specific configuration and processing described above and shown in the figures. For the sake of simplicity, a detailed description of the known method is omitted here. In the above embodiments, several specific steps are described and shown as examples. However, the method process of the present application is not limited to the specific steps described and shown, and those skilled in the art can make various changes, modifications and additions, or change the order between the steps after understanding the spirit of the present application.

[0300] The functional blocks shown in the structural block diagram described above can be implemented as hardware, software, firmware or a combination thereof. When implemented in hardware, it can be, for example, an electronic circuit, an application-specific integrated circuit (Application Specific Integrated Circuit, ASIC), appropriate firmware, plug-in, function card, etc. When implemented in software, the elements of the present application are programs or code segments used to perform the required tasks. The program or code segment can be stored in a machine-readable medium, or transmitted on a transmission medium or communication link by a data signal carried in a carrier. "Machine-readable medium" can include any medium capable of storing or transmitting information. Examples of machine-readable media include electronic circuits, semiconductor memory devices, read-only memory (Read-Only Memory, ROM), flash memory, erasable read-only memory (Erasable ReadOnly Memory, EROM), floppy disks, compact disc read-only memory (Compact Disc Read-Only Memory, CD-ROM), optical discs, hard disks, optical fiber media, radio frequency (Radio Frequency, RF) links, etc. The code segment can be downloaded via a computer network such as the Internet, an intranet, etc.

[0301] It should also be noted that the exemplary embodiments mentioned in this application describe some methods or systems based on a series of steps or devices. However, this application is not limited to the order of the above steps, that is, the steps can be performed in the order mentioned in the embodiment, or in a different order from the embodiment, or several steps can be performed simultaneously.

[0302] Aspects of the present disclosure are described above with reference to the flowchart and / or block diagram of the method, device (system) and computer program product according to the embodiment of the present disclosure. It should be understood that each box in the flowchart and / or block diagram and the combination of each box in the flowchart and / or block diagram can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device to produce a machine so that these instructions executed by the processor of the computer or other programmable data processing device enable the implementation of the function / action specified in one or more boxes of the flowchart and / or block diagram. Such a processor can be, but is not limited to, a general-purpose processor, a special-purpose processor, a special application processor, or a field programmable logic circuit. It can also be understood that each box in the block diagram and / or flowchart and the combination of boxes in the block diagram and / or flowchart can also be implemented by dedicated hardware that performs a specified function or action, or can be implemented by a combination of dedicated hardware and computer instructions.

[0303] The above are only specific implementation methods of the present application. Those skilled in the art can clearly understand that for the convenience and simplicity of description, the specific working processes of the systems, modules and units described above can refer to the corresponding processes in the aforementioned method embodiments, and will not be repeated here. It should be understood that the protection scope of the present application is not limited to this. Any technician familiar with the technical field can easily think of various equivalent modifications or replacements within the technical scope disclosed in this application, and these modifications or replacements should be included in the protection scope of this application.

Claims

1. A method for determining attribute information, characterized in that: include: Acquire multiple satisfaction evaluation information of a target time period and the weight of satisfaction evaluation attribute information in each satisfaction evaluation information; Select m target satisfaction evaluation attribute information with larger weights from the satisfaction evaluation attribute information, where m is a positive integer; Obtaining the correlation accuracy of each target satisfaction evaluation attribute information; According to the weight and the associated accuracy of each target satisfaction evaluation attribute information, determine the target weight of each target satisfaction evaluation attribute information according to a preset relationship, wherein the preset relationship includes the relationship between the weight, the associated accuracy and the target weight of the preset satisfaction evaluation attribute information, and the preset satisfaction evaluation attribute information includes the m target satisfaction evaluation attribute information; A preset number of target satisfaction evaluation attribute information with a larger target weight is determined as key satisfaction evaluation attribute information.

2. The method for determining attribute information according to claim 1, characterized in that: The weight includes a first sub-weight, a second sub-weight and a third sub-weight; The step of selecting m target satisfaction evaluation attribute information with larger weights from the satisfaction evaluation attribute information includes: From the satisfaction evaluation attribute information, select k first satisfaction evaluation attribute information with larger weights in the first sub-weights, l second satisfaction evaluation attribute information with larger weights in the second sub-weights, and n third satisfaction evaluation attribute information with larger weights in the third sub-weights; Combining the first satisfaction evaluation attribute information, the second satisfaction evaluation attribute information and the third satisfaction evaluation attribute information to obtain m target satisfaction evaluation attribute information; Among them, the sum of the first sub-weights of the k first satisfaction evaluation attribute information, the sum of the second sub-weights of the l second satisfaction evaluation attribute information, and the sum of the third sub-weights of the n third satisfaction evaluation attribute information are all not less than the first threshold; m≤k+l+n.

3. The method for determining attribute information according to claim 2, characterized in that: Obtaining the weight of the satisfaction evaluation attribute information in each of the satisfaction evaluation information includes: Obtain the frequency of the satisfaction evaluation attribute information in each satisfaction evaluation information, the evaluation level of the satisfaction evaluation attribute information in each satisfaction evaluation information, and the probability of the satisfaction evaluation attribute information in the satisfaction evaluation information of each satisfaction level; Determining, according to the frequency of the satisfaction evaluation attribute information, the proportion of the satisfaction evaluation attribute information in each of the satisfaction evaluation information; determining the entropy of the satisfaction evaluation attribute information according to the proportion of the satisfaction evaluation attribute information in each of the satisfaction evaluation information; Determining a first sub-weight of the satisfaction evaluation attribute information according to the entropy of the satisfaction evaluation attribute information; Determining the importance index of the satisfaction evaluation attribute information according to the evaluation level of the satisfaction evaluation attribute information in each of the satisfaction evaluation information and the number of the plurality of satisfaction evaluation attribute information; Normalizing the importance index of the satisfaction evaluation attribute information to obtain a second sub-weight of the satisfaction evaluation attribute information; The third sub-weight of the satisfaction evaluation attribute information is determined according to the probability of the satisfaction evaluation attribute information in the satisfaction evaluation information of each satisfaction level.

4. The method for determining attribute information according to claim 1, characterized in that: Before obtaining the correlation accuracy of each target satisfaction evaluation attribute information, the method further includes: Randomly selecting multiple first target satisfaction evaluation attribute information from the m target satisfaction evaluation attribute information; For each selection of a plurality of first target satisfaction evaluation attribute information, a first satisfaction prediction model is constructed using the plurality of first target satisfaction evaluation attribute information respectively, and the plurality of satisfaction evaluation information is input into the first satisfaction prediction model, and the satisfaction levels corresponding to the plurality of satisfaction evaluation information are predicted using the first satisfaction classification rules corresponding to the plurality of first target satisfaction evaluation attribute information in the first satisfaction prediction model, so as to obtain predicted satisfaction levels corresponding to the plurality of satisfaction evaluation information; Determining the accuracy of the first satisfaction prediction model according to the predicted satisfaction levels and the actual satisfaction levels corresponding to the plurality of satisfaction evaluation information; Calculating the average of the accuracy of the first satisfaction prediction model to obtain an average accuracy; According to the average accuracy and the number of the first target satisfaction evaluation attribute information, the associated accuracy of each target satisfaction evaluation attribute information is determined.

5. The method for determining attribute information according to claim 1, characterized in that: Before obtaining the weight of the satisfaction evaluation attribute information in each satisfaction evaluation information, the method further includes: Segmenting the plurality of satisfaction evaluation information to obtain a plurality of target segmentations; Deleting the segmentations that are consistent with the preset stop words from the multiple target segmentations to obtain multiple first segmentations; The segmentations in the first segmentations whose appearance frequency in the plurality of satisfaction evaluation information is less than a second threshold are deleted to obtain the satisfaction evaluation attribute information.

6. The method for determining attribute information according to claim 5, characterized in that: The step of deleting the segmentations in the first segmentations whose occurrence frequency in the plurality of satisfaction evaluation information is less than a second threshold to obtain the satisfaction evaluation attribute information includes: Deleting the words in the first participles whose appearance frequency in the plurality of satisfaction evaluation information is less than a second threshold, to obtain a plurality of second participles; For each target second participle in the plurality of second participles, calculating the similarity between the target second participle and a third participle, where the third participle is a participle in the plurality of second participles except the target second participle; From the plurality of second participles, third participles whose similarity to the target second participle is greater than a third threshold are deleted to obtain satisfaction evaluation attribute information.

7. The method for determining attribute information according to claim 6, characterized in that: The step of deleting, from the plurality of second participles, third participles whose similarity to the target second participle is greater than a third threshold value to obtain the satisfaction evaluation attribute information includes: From the plurality of second participles, delete the third participles whose similarity with the target second participle is greater than a third threshold, to obtain a plurality of fourth participles; A fourth participle whose appearance frequency in a plurality of preset words is greater than a fourth threshold is selected to obtain satisfaction evaluation attribute information.

8. The method for determining attribute information according to any one of claims 5 to 7, characterized in that: The method further comprises: Comparing each of the satisfaction evaluation information with the satisfaction evaluation attribute information to obtain a comparison result; Generating characteristic information of each satisfaction evaluation information according to the comparison result; The characteristic information is input into a second satisfaction prediction model, and the satisfaction levels corresponding to the multiple satisfaction evaluation information are predicted according to the second satisfaction classification rule in the second satisfaction prediction model to obtain the satisfaction levels corresponding to the multiple satisfaction evaluation information.

9. The method for determining attribute information according to claim 8, characterized in that: Before inputting the characteristic information into the second satisfaction prediction model, the method further includes: Acquire multiple satisfaction evaluation information samples and actual satisfaction levels corresponding to the satisfaction evaluation information samples; Segmenting the multiple satisfaction evaluation information samples to obtain multiple satisfaction evaluation attribute information samples; Comparing each of the satisfaction evaluation information samples with the multiple satisfaction evaluation attribute information samples to obtain a sample comparison result; Generating a characteristic information sample of each satisfaction evaluation information according to the sample comparison result; Inputting the characteristic information sample into the initial second satisfaction prediction model, predicting the satisfaction levels corresponding to the multiple satisfaction evaluation information samples according to the initial second satisfaction classification rules in the initial second satisfaction prediction model, and obtaining the sample predicted satisfaction levels corresponding to the multiple satisfaction evaluation information samples; Determining the accuracy of the initial second satisfaction prediction model according to the sample predicted satisfaction levels corresponding to the multiple satisfaction evaluation information samples and the actual satisfaction levels of the multiple satisfaction evaluation information samples; Update the initial second satisfaction classification rules in the initial second satisfaction prediction model, predict the satisfaction levels corresponding to the multiple satisfaction evaluation information samples according to the updated initial second satisfaction classification rules, and obtain updated sample predicted satisfaction levels; update the accuracy of the initial second satisfaction prediction model according to the updated sample predicted satisfaction levels and the actual satisfaction levels of the multiple satisfaction evaluation information samples; when the accuracy of the initial second satisfaction prediction model is greater than a fourth threshold, obtain the second satisfaction prediction model, which includes the second satisfaction classification rules.

10. A device for determining attribute information, characterized in that: include: An acquisition module, used to acquire multiple satisfaction evaluation information of a target time period and the weight of the satisfaction evaluation attribute information in each satisfaction evaluation information; A selection module, used for selecting m target satisfaction evaluation attribute information with larger weights from the satisfaction evaluation attribute information, where m is a positive integer; The acquisition module is further used to obtain the correlation accuracy of each target satisfaction evaluation attribute information; A determination module, configured to determine the target weight of each target satisfaction evaluation attribute information according to a preset relationship based on the weight and the associated accuracy of each target satisfaction evaluation attribute information, wherein the preset relationship includes the relationship between the weight, the associated accuracy and the target weight of the preset satisfaction evaluation attribute information, and the preset satisfaction evaluation attribute information includes the m target satisfaction evaluation attribute information; The determination module is further used to determine a preset number of target satisfaction evaluation attribute information with a larger target weight as key satisfaction evaluation attribute information.

11. An electronic device, characterized in that: The device comprises: a processor and a memory storing computer program instructions; When the processor executes the computer program instructions, the method for determining the attribute information according to any one of claims 1 to 9 is implemented.

12. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer program instructions, and when the computer program instructions are executed by a processor, the method for determining attribute information according to any one of claims 1 to 9 is implemented.

13. A computer program product, characterized in that When the instructions in the computer program product are executed by a processor of an electronic device, the electronic device is enabled to execute the method for determining attribute information as described in any one of claims 1 to 9.