A scoring method for user satisfaction with products based on fuzzy numbers

Through the fuzzy number scoring method, the parameters of lower limit, center and concentration are used to aggregate user scores, which solves the problem that the traditional scoring mechanism cannot accurately reflect the user's true thoughts, and achieves a more accurate user satisfaction assessment.

CN114579633BActive Publication Date: 2024-11-26BEIJING NORMAL UNIVERSITY
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202210198093.8
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2022-03-02
Publication Date
2024-11-26
Estimated Expiration
2042-03-02

AI Technical Summary

Technical Problem

In the prior art, when evaluating user satisfaction in online shopping software, the traditional precise scoring mechanism cannot effectively reflect the user's true thoughts, resulting in inaccurate evaluation results, especially when users accumulate ratings in multiple aspects.

Method used

Using the user's product satisfaction rating method based on fuzzy numbers, the user's rating is described as a fuzzy number containing the lower limit, center, upper limit, left and right concentrations, and construct a new fuzzy number through a weighted average aggregation method. ,以更好地反映用户的主观评分。

Benefits of technology

This method can better describe the user's real ratings, provide more accurate and comprehensive psychological intentions, and avoid the loss of information during rating aggregation in traditional methods.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN114579633B_ABST
    Figure CN114579633B_ABST
Patent Text Reader

Abstract

The present invention discloses a method for scoring user satisfaction with a product based on fuzzy numbers, comprising the following steps: S1, describing the scores of all users participating in the scoring with a fuzzy number, wherein the parameters of the fuzzy number include a lower limit a, a center b, an upper limit c, a left concentration μ ‑ and the right side concentration μ + ; S2. Aggregate the fuzzy numbers. The aggregation method is to obtain new parameters by taking weighted average of each parameter, and use the new parameters to determine a new fuzzy number. The advantages are: introducing fuzzy numbers to describe the user's comprehensive evaluation of the product. Compared with the traditional use of exact numbers, using fuzzy numbers to describe the user's rating can better describe the subjective and real rating.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and particularly relates to a method for scoring the satisfaction of users with commodities based on fuzzy numbers. Background Art

[0002] An element has only two relationships with a classical set, namely "belong to" and "not belong to". However, some concepts are fuzzy and do not conform to the concept of "set". Therefore, people defined the "fuzzy set": the relationship between an element and a fuzzy set is "belong to with a membership degree of m", where m is a real number on [0, 1], representing the degree to which the element belongs to the set, called the "membership degree". m = 0 means completely not belong to, and m = 1 means completely belong to. In this sense, the classical set is a special case of the fuzzy set, and its special point is that the membership degree m can only take 0 or 1. In a fuzzy set, the corresponding relationship between each element and its membership degree to the set is called the "membership function".

[0003] The role of the fuzzy set is to reflect the objective world while adding a certain degree of subjectivity. Many fuzzy concepts, such as "tall, short, fat, thin", etc., are not completely objective but carry a certain degree of subjectivity. The field of artificial intelligence needs to interact with people and often encounters concepts with subjectivity. Therefore, it is necessary to introduce the fuzzy set.

[0004] A concept like "about 3" is called a fuzzy number. The essence of a fuzzy number is a fuzzy set, and each specific number corresponds to a membership degree (for example, in a specific context, 3.8 is barely "about 3", and its membership degree is set to 0.5), thus constituting a membership function μ(x). From this, we can determine a fuzzy number by the membership function μ(x). The membership function of the fuzzy number "about b" can have many cases, and they need to meet the following conditions:

[0005] ① It is monotonically increasing when x < b (i.e., μ'(x) ≥ 0) and monotonically decreasing when x > b (i.e., μ'(x) ≤ 0). ② μ(b) = 1; when x approaches negative infinity or positive infinity, μ(x) = 0.

[0006] Previous researchers constructed various fuzzy numbers. The simplest one is the "triangular fuzzy number", and its membership function form is as shown in the appendix Figure 1 shown. If the set with a membership degree of 1 is not a single point but an interval, it is called a "trapezoidal fuzzy number", and its membership function form is as shown in the appendix Figure 2 shown. If a given membership function passes through a fixed point (i.e., given the membership degree corresponding to a certain x), then this point is called a "control point". The simplest form of a fuzzy number with a control point is as shown in the appendix Figure 3 shown.

[0007] The commonly used fuzzy number calculation method is defined as follows: the membership function of the fuzzy number is cut horizontally into countless slices (such as the attached Figure 4 As shown in Figure 1, each slice is a classical set, corresponding to a membership level λ (the corresponding slice is called a "λ-cut set"). These classical sets are operated accordingly and then superimposed according to the membership level to obtain the result of fuzzy number calculation.

[0008] The function spaces corresponding to some existing fuzzy number and operation construction methods are not clear enough, and the theoretical foundation is not solid enough, which leaves room for problems in future applications. Therefore, we need to establish a clear function space and make this function space complete for algebraic operations.

[0009] In the process of investigating user satisfaction of online shopping software, after collecting questionnaires, the user satisfaction is usually evaluated by using conventional weighted average calculation and other methods to aggregate data. The disadvantage of this method is that aggregating a user's scores in multiple aspects into one value will inevitably lose some information, resulting in the evaluation results not being able to reflect the actual situation well.

[0010] For example, two users both gave a 4 out of 5 rating to the same product, but one of them thought that 4 was a bit high, but 3 was too low, so he chose 4 after hesitating. The other user hesitated between 4 and 5 and chose 4. The two gave the same rating but had different intentions. This shows that the traditional precise rating mechanism cannot reasonably express the user's true thoughts due to the limitation of rating accuracy. Even a user may give a high rating when he is in a good mood and a low rating when he is in a bad mood. Therefore, for each user, fuzzy numbers can better reflect his true rating than precise numbers. Summary of the invention

[0011] The purpose of the present invention is to provide a method for scoring user satisfaction with a product based on fuzzy numbers, thereby solving the above-mentioned problems existing in the prior art.

[0012] In order to achieve the above object, the technical solution adopted by the present invention is as follows:

[0013] A method for scoring user satisfaction with a product based on fuzzy numbers includes the following steps:

[0014] S1. Describe the ratings of all users who participated in the rating with a fuzzy number. The parameters of this fuzzy number include the lower limit a, the center b, the upper limit c, and the left concentration μ. - and the right side concentration μ + ;

[0015] S2. Aggregate the fuzzy numbers by taking the weighted average of the parameters to obtain new parameters, and use the new parameters to determine a new fuzzy number.

[0016] Preferably, step S1 specifically includes the following contents:

[0017] S11. Obtain a user's satisfaction rating on several aspects of the product through a questionnaire survey; obtain the corresponding lower limit a, center b, upper limit c, and left concentration μ based on the rating. - Or the left control point (x1, y1), the right concentration μ + or right control point (x2,y2);

[0018] S12: If the left control point or the right control point is obtained in step S11 instead of the concentration, the left concentration μ is determined according to the left control point and the right control point. - and / or right side concentration μ + ;

[0019] S13, based on the obtained lower limit a, center b, upper limit c, left concentration μ - , right side concentration μ + Construct fuzzy numbers.

[0020] Preferably, step S11 specifically includes:

[0021] Questionnaire Overview: Each question has a full score of 10 points, and the next question will appear after each question is answered;

[0022] Problem display:

[0023] Question 1: What score do you think is appropriate?

[0024] Corresponding parameters: b;

[0025] Optional items: 0-10, precision is 0.5;

[0026] Question 2: What is the minimum score you think should be given?

[0027] Corresponding parameters: a;

[0028] Optional value: 0-10, precision is 1;

[0029] Question 3: What is the maximum score you think can be given?

[0030] Corresponding parameters: c;

[0031] Optional value: 0-10, precision is 1;

[0032] Question 4: What score do you think is appropriate for x1?

[0033] Corresponding parameters: left control point (x1, y1);

[0034] x1 is a randomly generated number on (a,b) with a range of 0-10 and a precision of 0.5; y1 is the value corresponding to the user's answer;

[0035] Options: Unsuitable, Average, and Relatively Suitable; the corresponding value for Unsuitable is 0.3, the corresponding value for Average is 0.5, and the corresponding value for Relatively Suitable is 0.7;

[0036] Question 5: What score do you think is appropriate for x2?

[0037] Corresponding parameters: right control point (x2, y2);

[0038] x2 is a randomly generated number on (a,b) with a range of 0-10 and a precision of 0.5; y2 is the value corresponding to the user's answer;

[0039] Options: Unsuitable, Average, and Relatively Suitable; the corresponding value for Unsuitable is 0.3, the corresponding value for Average is 0.5, and the corresponding value for Relatively Suitable is 0.7;

[0040] When the following situations occur, the left side concentration μ is directly obtained according to the corresponding situation - Or right side concentration μ + :

[0041] (1) If the answer to question 1 is 0, omit questions 2 and 4 and directly set a = 0, μ - =0;

[0042] (2) If the answer to question 1 is 10, omit questions 3 and 5 and directly set c = 10, μ + =0;

[0043] (3) If ba = 0.5 or 0, skip question 4 and directly set μ - =0;

[0044] (4) If cb = 0.5 or 0, skip question 5 and directly set μ + =0.

[0045] (5) If a>b or b>c, the questionnaire is considered invalid.

[0046] Preferably, in step S12, the left concentration μ is determined according to the left control point and the right control point. - and / or right side concentration μ + The calculation formula is as follows,

[0047]

[0048] q(y)=F -1 (y)

[0049] μ - =h(x1;a,b)-q(y1)

[0050] μ + =h(x2;c,b)-q(y2).

[0051] Preferably, the membership function of the fuzzy number constructed in step S13 is:

[0052]

[0053] in,

[0054] Preferably, step S2 is specifically as follows: let the kth fuzzy number constructed in step S1 be The following formula is used to calculate the weighted average of each parameter, and then a new fuzzy number is constructed:

[0055] a=∑w k a k

[0056] b=∑w k b k

[0057] c=∑w k c k

[0058]

[0059]

[0060] Where k = 1, 2, ..., n, n is the total number of users participating in the rating; a k is the minimum value of the k-th user's rating; b k is the average value of the kth user's rating, c k The maximum value of k user ratings, The left concentration of k user ratings, is the right-hand concentration of the k-th user's rating; w k is the weight of the kth user.

[0061] The beneficial effects of the present invention are as follows: introducing fuzzy numbers to describe users' comprehensive evaluation of commodities, and using fuzzy numbers to describe users' ratings can better describe subjective and real ratings compared with traditional use of precise numbers. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] Figure 1 It is a schematic diagram of the triangular fuzzy number form in the prior art;

[0063] Figure 2 It is a schematic diagram of the trapezoidal fuzzy number form in the prior art;

[0064] Figure 3 It is the simplest form of the fuzzy number of the control point in the prior art;

[0065] Figure 4 It is a schematic diagram of a horizontal slice of a fuzzy number membership function in the prior art;

[0066] Figure 5 is a schematic flow chart of a method in an embodiment of the present invention;

[0067] Figure 6 Schematic diagram of determining a membership function through control points in an embodiment of the present invention. DETAILED DESCRIPTION

[0068] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings. It should be understood that the specific implementation methods described herein are only used to explain the present invention and are not used to limit the present invention.

[0069] like Figure 5 As shown, in this embodiment, a method for scoring user satisfaction with a product based on fuzzy numbers is provided, comprising the following steps:

[0070] S1. Describe the ratings of all users who participated in the rating with a fuzzy number. The parameters of this fuzzy number include the lower limit a, the center b, the upper limit c, and the left concentration μ. - and the right side concentration μ + ;

[0071] S2. Aggregate the fuzzy numbers by taking the weighted average of the parameters to obtain new parameters, and use the new parameters to determine a new fuzzy number.

[0072] In this embodiment, the fuzzy number The membership function f(x) is composed of a monotonically increasing function f on the interval (a, b). - (x) and a monotonically decreasing function f on the interval (b,c) + (x) and satisfies

[0073]

[0074] where f + (x) and f - (x) is composed of nonlinear functions h(x) and p(x). p(x) is the probability density function of the normal distribution with parameter (μ,1), and the analytical expression is The composition is:

[0075]

[0076] In summary, the membership function f(x) contains parameters a, b, c, μ - ,μ + , so it is recorded as f(x; a, b, c, μ - ,μ + ). Its corresponding fuzzy number is recorded as <a,b,c;μ - ,μ + >.

[0077] If a fuzzy number Only the parameters a, b, c and the control points (x1, y1) and (x2, y2) of the membership function are given (both in the increasing and decreasing segments, respectively), but the concentration degree μ on both sides is not given. - ,μ + , then the concentration degree on both sides can be calculated through the control points. The solution is as follows (taking the left control point (x1, y1) as an example):

[0078] Solve the left side concentration μ by equation f(x1=y1) - , which is solved as μ - =h(x1;a,b)-q(y1), where q is the quantile function of the standard normal distribution, i.e. q(y)=F -1 (y), Solve the right side concentration μ in the same way + =h(x2;c,b)-q(y2). An example of determining the membership function through control points is Figure 6 shown.

[0079] Step S1 specifically includes the following contents:

[0080] S11. Obtain a user's satisfaction rating on several aspects of the product through a questionnaire survey; obtain the corresponding lower limit a, center b, upper limit c, and left concentration μ based on the rating. - Or the left control point (x1, y1), the right concentration μ + or right control point (x2,y2);

[0081] Specifically, when evaluating user (buyer) satisfaction with a product in an online shopping software, user satisfaction data is first collected through a user feedback questionnaire. The questionnaire contains several aspects (dimensions), such as product quality, delivery speed, mailing service quality, etc. (which may vary depending on the product). The specific content of the questionnaire survey on user satisfaction with relevant aspects of a product is as follows:

[0082] Questionnaire Overview: Displayed in the form of a web page, each question has a full score of 10 points, and the next question will appear after each question is answered;

[0083] Problem display:

[0084] Question 1: What score do you think is appropriate?

[0085] Corresponding parameters: b;

[0086] Optional items: 0-10, precision is 0.5;

[0087] Question 2: What is the minimum score you think should be given?

[0088] Corresponding parameters: a;

[0089] Optional value: 0-10, precision is 1;

[0090] Question 3: What is the maximum score you think can be given?

[0091] Corresponding parameters: c;

[0092] Optional value: 0-10, precision is 1;

[0093] Question 4: What score do you think is appropriate for x1?

[0094] Corresponding parameters: left control point (x1, y1);

[0095] x1 is a randomly generated number on (a,b) with a range of 0-10 and a precision of 0.5; y1 is the value corresponding to the user's answer;

[0096] Options: Unsuitable, Average, and Relatively Suitable; the corresponding value for Unsuitable is 0.3, the corresponding value for Average is 0.5, and the corresponding value for Relatively Suitable is 0.7;

[0097] Question 5: What score do you think is appropriate for x2?

[0098] Corresponding parameters: right control point (x2, y2);

[0099] x2 is a randomly generated number on (a,b) with a range of 0-10 and a precision of 0.5; y1 is the value corresponding to the user's answer;

[0100] Options: Unsuitable, Average, and Relatively Suitable; the corresponding value for Unsuitable is 0.3, the corresponding value for Average is 0.5, and the corresponding value for Relatively Suitable is 0.7;

[0101] When the following situations occur, the left side concentration μ is directly obtained according to the corresponding situation - Or right side concentration μ + :

[0102] (1) If the answer to question 1 is 0, omit questions 2 and 4 and directly set a = 0, μ - =0;

[0103] (2) If the answer to question 1 is 10, omit questions 3 and 5 and directly set c = 10, μ + =0;

[0104] (3) If ba = 0.5 or 0, skip question 4 and directly set μ - =0;

[0105] (4) If cb = 0.5 or 0, skip question 5 and directly set μ + =0.

[0106] (5) If a>b or b>c, the questionnaire is considered invalid.

[0107] S12: If the left control point or the right control point is obtained in step S11 instead of the concentration, the left concentration μ is determined according to the left control point and the right control point. - and / or right side concentration μ + ; The specific calculation formula is as follows:

[0108]

[0109] q(y)=F -1 (y)

[0110] μ - =h(x1;a,b)-q(y1)

[0111] μ + =h(x2;c,b)-q(y2).

[0112] S13, based on the obtained lower limit a, center b, upper limit c, left concentration μ - , right side concentration μ + Construct a fuzzy number. The membership function of the fuzzy number is:

[0113]

[0114] in,

[0115] The fuzzy number constructed using this <a,b,c;μ - ,μ + > to describe a user's opinion.

[0116] In this embodiment, step S2 is specifically as follows: assuming that the kth fuzzy number constructed in step S1 is The following formula is used to calculate the weighted average of each parameter, and then a new fuzzy number is constructed:

[0117] a=∑w k a k

[0118] b=∑w k b k

[0119] c=∑w k c k

[0120]

[0121]

[0122] Where k = 1, 2, ..., n, n is the total number of users participating in the rating; a k is the minimum value of the k-th user's rating; b k is the average value of the kth user's rating, c k The maximum value of k user ratings, The left concentration of k user ratings, is the right-hand concentration of the k-th user's rating; w k is the weight of the kth user.

[0123] User weight can be obtained in many ways, such as the level obtained by the user through points in the software, or the consumption amount, etc. Many softwares have already developed mature methods for generating user weights. It is even possible to directly set the weight of each user to 1, that is, degenerate to the traditional average.

[0124] The new fuzzy number finally constructed can be used as the final score of user satisfaction with the product and then displayed to the public. On the one hand, suppliers can make supply adjustments based on the score, and on the other hand, customers can estimate the quality of the product based on the score.

[0125] By adopting the above technical solution disclosed in the present invention, the following beneficial effects are obtained:

[0126] The present invention provides a method for scoring user satisfaction with a product based on fuzzy numbers. The method introduces fuzzy numbers to describe the user's comprehensive evaluation of the product. Compared with the traditional use of exact numbers, the use of fuzzy numbers to describe the user's rating can better describe the subjective and real rating.

[0127] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be considered as the scope of protection of the present invention.

Claims

1. A method for scoring user satisfaction with a product based on fuzzy numbers, characterized by: The following steps are included: S1. Describe the ratings of all users who participated in the rating with a fuzzy number. The parameters of this fuzzy number include the lower limit a, the center b, the upper limit c, and the left concentration μ. - and the right side concentration μ + ; S2, aggregate the fuzzy numbers by taking the weighted average of the parameters to obtain new parameters, and use the new parameters to determine a new fuzzy number; Step S1 specifically includes the following contents: S11. Obtain a user's satisfaction rating on several aspects of the product through a questionnaire survey; obtain the corresponding lower limit a, center b, upper limit c, and left concentration μ based on the rating. - Or the left control point (x1, y1), the right concentration μ + or right control point (x2,y2); S12: If the left control point or the right control point is obtained in step S11 instead of the concentration, the left concentration μ is determined according to the left control point and the right control point. - and / or right side concentration μ + ; In step S12, the left concentration μ is determined according to the left control point and the right control point. - and / or right side concentration μ + The calculation formula is as follows, q(y)=F -1 (y) m - =h(x1;a,b)-q(y1) μ + =h(x2;c,b)-q(y2); S13, based on the obtained lower limit a, center b, upper limit c, left concentration μ - , right side concentration μ + Construct fuzzy numbers; The membership function of the fuzzy number constructed in step S13 is: In step S12 and step S13, 2. The method for evaluating user satisfaction with a product based on fuzzy numbers according to claim 1, characterized in that: Step S11 specifically includes: Questionnaire Overview: Each question has a full score of 10 points, and the next question will appear after each question is answered; Problem display: Question 1: What score do you think is appropriate? Corresponding parameters: b; Optional items: 0-10, precision is 0.5; Question 2: What is the minimum score you think should be given? Corresponding parameters: a; Optional value: 0-10, precision is 1; Question 3: What is the maximum score you think can be given? Corresponding parameters: c; Optional value: 0-10, precision is 1; Question 4: What score do you think is appropriate for x1? Corresponding parameters: left control point (x1, y1); x1 is a randomly generated number on (a,b) with a range of 0-10 and a precision of 0.5; y1 is the value corresponding to the user's answer; Options: Unsuitable, Average, and Relatively Suitable; the corresponding value for Unsuitable is 0.3, the corresponding value for Average is 0.5, and the corresponding value for Relatively Suitable is 0.7; Question 5: What score do you think is appropriate for x2? Corresponding parameters: right control point (x2, y2); x2 is a randomly generated number on (a,b) with a range of 0-10 and a precision of 0.5; y2 is the value corresponding to the user's answer; Options: Unsuitable, Average, and Relatively Suitable; the corresponding value for Unsuitable is 0.3, the corresponding value for Average is 0.5, and the corresponding value for Relatively Suitable is 0.7; When the following situations occur, the left side concentration μ is directly obtained according to the corresponding situation - Or right side concentration μ + : (1) If the answer to question 1 is 0, omit questions 2 and 4 and directly set a = 0, μ - =0; (2) If the answer to question 1 is 10, omit questions 3 and 5 and directly set c = 10, μ + =0; (3) If ba = 0.5 or 0, skip question 4 and directly set μ - =0; (4) If cb = 0.5 or 0, skip question 5 and directly set μ + =0; (5) If a>b or b>c, the questionnaire is considered invalid.

3. The method for evaluating user satisfaction with a product based on fuzzy numbers according to claim 1, characterized in that: Specifically, step S2 is: let the kth fuzzy number constructed in step S1 be The following formula is used to calculate the weighted average of each parameter, and then a new fuzzy number is constructed: a=∑w k a k b=∑w k b k c=∑w k c k Where k = 1, 2, ..., n, n is the total number of users participating in the rating; a k is the minimum value of the k-th user's rating; b k is the average value of the kth user's rating, c k is the maximum value of the k-th user's rating, is the left concentration of the k-th user rating, is the right-hand concentration of the k-th user's rating; w k is the weight of the kth user.