Decision method and device, terminal equipment and computer readable storage medium

Abstract

The invention provides a decision-making method and device, terminal equipment and a computer readable storage medium. The method comprises the steps of obtaining evaluation information corresponding to each alternative scheme; determining evaluation result information of each piece of evaluation information for the respective corresponding alternative scheme; and determining at least one target scheme from the alternative schemes according to the evaluation result information. According to the decision-making method provided by the invention, the evaluation information of the decision maker for each alternative scheme is obtained, and the evaluation result information of each decision maker for different alternative schemes is analyzed from each evaluation information, so that the selection tendency of each decision maker for different alternative schemes can be determined; therefore, each decision maker can be effectively helped to determine an effective target scheme.

Description

Technical Field

[0001] The present application relates to the field of computer technology, and in particular to a decision-making method, apparatus, terminal device and computer-readable storage medium. Background Art

[0002] Large group decision-making refers to the process in which many decision-makers participate (generally referring to group decision-making of more than one person), express preferences for alternative plans based on the decision information provided, and aggregate many individual judgments into large group judgments according to specific rules to rank the plans or select the best plan from multiple alternative plans.

[0003] In the process of large group decision analysis, due to individual differences between people, they are not on the same standard when making decisions. Some decision makers will make decisions optimistically, some will make decisions pessimistically, and some cannot use accurate data to measure their decision ideas. Therefore, there is currently no relevant method to solve the problem of how to make the expression standards of decision makers consistent to the greatest extent. Summary of the invention

[0004] The present application relates to a decision-making method, which helps users make decisions on solutions by determining the inclinations of each user.

[0005] In a first aspect, the present application provides a decision-making method, the method comprising:

[0006] Obtain the evaluation information corresponding to each alternative plan;

[0007] Determining evaluation result information of each evaluation information for the corresponding alternative solution;

[0008] According to the evaluation result information, at least one target solution is determined from the alternative solutions.

[0009] In a second aspect, the present application further provides a decision-making device, the device comprising:

[0010] An acquisition module is used to obtain evaluation information corresponding to each alternative plan;

[0011] A determination module, used to determine the evaluation result information of each evaluation information for the corresponding alternative solution;

[0012] The decision module is used to determine at least one target solution from the alternative solutions according to the evaluation result information.

[0013] In a third aspect, the present application also provides a terminal device, characterized in that the terminal device includes a processor, a memory, and a computer program stored in the memory and executable on the processor, and the processor executes the computer program to implement the steps in any one of the decision-making methods.

[0014] In a fourth aspect, the present application further provides a computer-readable storage medium, on which a computer program is stored, and the computer program is executed by a processor to implement the steps in any one of the decision-making methods.

[0015] The decision-making method provided in the present application obtains the evaluation information of the decision makers for each alternative plan, and then analyzes the evaluation result information of each decision maker for different alternative plans from each evaluation information, so as to determine the selection tendency of each decision maker for different alternative plans, thereby effectively helping each decision maker to determine an effective target plan. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative work.

[0017] Figure 1 is a scenario diagram of a decision-making system provided in an embodiment of the present application;

[0018] Figure 2 This is a flowchart of an embodiment of the decision-making method in the embodiment of the present application;

[0019] Figure 3 This is a flowchart of an embodiment of the decision-making method in the embodiment of the present application;

[0020] Figure 4 It is a functional module diagram of a decision-making device in an embodiment of the present application;

[0021] Figure 5 It is a schematic diagram of the structure of the terminal device in the embodiment of the present application. DETAILED DESCRIPTION

[0022] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of this application.

[0023] In the description of the present application, it should be understood that the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the feature. In the description of the present application, "plurality" means two or more, unless otherwise clearly and specifically defined.

[0024] In this application, the word "exemplary" is used to mean "used as an example, illustration or description". Any embodiment described as "exemplary" in this application is not necessarily to be construed as being more preferred or advantageous than other embodiments. At the same time, it is to be understood that in the specific implementation of this application, when user information, user data and other related data are involved, when the above embodiments of this application are applied to specific products or technologies, user permission or consent is required, and the collection, use and processing of relevant data shall comply with relevant laws, regulations and standards of relevant countries and regions.

[0025] In order to enable any person skilled in the art to implement and use the present application, the following description is provided. In the following description, details are listed for the purpose of explanation. It should be understood that those of ordinary skill in the art will recognize that the present application can be implemented without using these specific details. In other examples, known structures and processes will not be elaborated in detail to avoid unnecessary details that make the description of the present application obscure. Therefore, the present application is not intended to be limited to the embodiments shown, but is consistent with the widest range of principles and features disclosed in the present application.

[0026] The present application provides a decision-making method, apparatus, device and storage medium, which are described in detail below.

[0027] See also Figure 1 , Figure 1 Schematic diagram of a decision system provided in an embodiment of the present application. The decision system may include a terminal device 100 and a storage device 200. The storage device 200 may transmit data to the terminal device 100. Figure 1 The terminal device 100 in the embodiment can obtain the evaluation information for each alternative solution stored in the storage device 200 to execute the decision-making method in the present application.

[0028] In the embodiment of the present application, the terminal device 100 includes but is not limited to a desktop computer, a portable computer, a network server, a PDA (Personal Digital Assistant), a tablet computer, a wireless terminal device, an embedded device, etc.

[0029] In an embodiment of the present application, communication between the terminal device 100 and the storage device 200 can be achieved through any communication method, including but not limited to mobile communications based on the 3rd Generation Partnership Project (3GPP), Long Term Evolution (LTE), and Worldwide Interoperability for Microwave Access (WiMAX), or computer network communications based on the TCP / IP Protocol Suite (TCP / IP) and User Datagram Protocol (UDP).

[0030] It should be noted that Figure 1 The scenario diagram of the decision-making system shown is merely an example. The decision-making system and scenario described in the embodiment of the present application are intended to more clearly illustrate the technical solution of the embodiment of the present application, and do not constitute a limitation on the technical solution provided in the embodiment of the present application. A person of ordinary skill in the art can appreciate that with the evolution of the decision-making system and the emergence of new business scenarios, the technical solution provided in the embodiment of the present application is equally applicable to similar technical problems.

[0031] like Figure 2 As shown, Figure 2 This is a flow chart of an embodiment of a decision-making method in an embodiment of the present application. The decision-making method may include the following steps 201 to 203:

[0032] 201. Obtain evaluation information corresponding to each alternative plan.

[0033] The alternative solutions in the embodiments of the present application may be artificial solutions proposed for specific matters. It should be noted that in the embodiments of the present application, there is no need to obtain the specific solution content, only the solution number is required, so that the solution number can be output after the subsequent decision on the specific target solution. At the same time, in the embodiments of the present application, the evaluation information can be the decision maker's evaluation text for each alternative solution, such as: txt text, word text, and of course, it can also be the decision maker's handwritten text. If the decision maker's handwritten text is obtained, the user's handwritten text can be identified through image recognition and converted into txt text or word text.

[0034] In addition, in the embodiment of the present application, the evaluation information can be obtained according to the implementation of the above-mentioned embodiment, and the corresponding evaluation information can be obtained from the relevant storage medium or the cloud, wherein each evaluation information can have a mark of the corresponding scheme to help the decision-making device of the embodiment of the present application determine the corresponding relationship between the obtained evaluation information and the alternative scheme. Specifically, assuming that there are decision makers A and decision makers B, and there are alternative schemes 1 and 2, then the evaluation information of decision maker A on alternative scheme 1 is A1, the evaluation information of decision maker A on alternative scheme 2 is A2, the evaluation information of decision maker B on alternative scheme 1 is B1, and the evaluation information of decision maker B on alternative scheme 2 is B2. The marks of the evaluation information A1, A2, B1, and B2 can characterize which decision maker the evaluation information comes from and which alternative scheme it corresponds to. Of course, the decision makers can also be decision makers C, decision makers D, decision makers E, decision makers F, etc., and the alternative schemes can also be alternative schemes 2, alternative schemes 3, alternative schemes 4, alternative schemes 5, alternative schemes 6, etc., and the specific embodiments of the present application are not limited.

[0035] Of course, in the embodiments of the present application, the evaluation information can also be obtained in real time. For example: there is a meeting to decide on alternative plans, and the host of the meeting can input the number of the relevant alternative plan into the decision-making device. Afterwards, when each decision maker in the meeting has read the contents of each specific alternative plan, each decision maker can generate an evaluation information txt or word for each alternative plan, and transmit the file of each evaluation information to the meeting host. After the meeting host receives the file of each evaluation information, he inputs each evaluation information into the corresponding alternative plan, and the decision-making device can obtain the evaluation information corresponding to each alternative plan. Alternatively, it can also be handed over to each decision maker to upload the generated evaluation information to the decision-making device. The specific evaluation information corresponding to each alternative plan can be based on the actual situation, and the embodiments of the present application do not make specific limitations.

[0036] 202. Determine evaluation result information of each evaluation information for its corresponding alternative solution.

[0037] After obtaining the evaluation information corresponding to each alternative plan, the words in the evaluation information can be analyzed, and the evaluation result information of each decision maker for different alternative plans can be analyzed therefrom. Since the evaluation information comes from each decision maker, and according to the above embodiment, the evaluation information can contain the identity mark of the decision maker and the mark of the corresponding alternative plan. Therefore, when the evaluation result information of each decision maker for different alternative plans can be analyzed, the evaluation result information of each evaluation information for the corresponding alternative plan can also be determined in disguise.

[0038] In the embodiment of the present application, each evaluation information can be input into the BERT model to obtain the word vector of each evaluation information, E s =[e i1 , e i2 ,…,e in ]∈R N*demb Among them, R N*demb is the size of the real number space, used to describe the word vector E s Dimension. After obtaining the word vector in the evaluation information, the word vector can be compared with some preset reference word vectors with emotional words, for example, the similarity between the word vector in the evaluation information and the reference word vector is calculated. When the similarity is less than a certain threshold, it can be determined that the word vector in the evaluation information hits the reference word vector currently being compared. The more reference word vectors the word vector in the evaluation information hits, the higher the tendency of the current evaluation information for a certain target solution, the higher the score information, and the more the decision maker tends to the alternative solution corresponding to the current evaluation information. In this way, the tendency of each evaluation information for its corresponding alternative solution can be obtained. In addition, since each decision maker needs to generate an evaluation information for each alternative solution. Therefore, a score information can be determined based on the number of hits of the word vector of each evaluation information, thereby determining the alternative solution that a decision maker is most inclined to. For example, if we take the marking of the above embodiment as an example, assuming that decision maker A corresponds to evaluation information A1 and evaluation information A2, if the number of hits of evaluation information A1 is large, a higher score is obtained; and the number of hits of evaluation information A2 is small, and the score information is small, it proves that decision maker A is more inclined to alternative plan 1 than alternative plan 2. The inclination of each decision maker for different alternative plans in this step can be regarded as the evaluation result information of each evaluation information for the corresponding alternative plan.

[0039] 203. According to the evaluation result information, at least one target solution is determined from the alternative solutions.

[0040] After obtaining the evaluation result information of each evaluation information for different alternatives, the final target solution can be determined from each alternative solution according to each evaluation result information. For example: Assuming there are decision makers A and B, and there are alternatives 1 and 2, then the evaluation information of decision maker A for alternative 1 is A1, the evaluation information of decision maker A for alternative 2 is A2, the evaluation information of decision maker B for alternative 1 is B1, and the evaluation information of decision maker B for alternative 2 is B2. At this time, if the tendency represented by the evaluation result information of evaluation information A1 is greater than the tendency represented by the evaluation result information of evaluation information A2, it proves that decision maker A is more inclined to choose alternative 1. Similarly, if the tendency represented by the evaluation result information of evaluation information B1 is greater than the tendency represented by the evaluation result information of evaluation information B2, it proves that decision maker B is also more inclined to choose alternative 1. At this time, alternative 1 can be determined as the final target solution, and the decision-making device outputs the number 1.

[0041] However, it should be noted that if the tendency represented by the evaluation result information of evaluation information B1 is less than the tendency represented by the evaluation result information of evaluation information B2, it proves that decision maker B is more inclined to choose alternative plan 2. At this time, alternative plan 1 and alternative plan 2 are in a tie, and the decision-making device can prompt the decision maker whether to revise his evaluation information. If the decision maker does not upload new evaluation information within the specified time, a simulated voting result of each alternative plan can be generated; or, if the decision maker uploads new evaluation information within the specified time, the alternative plan is re-determined as the final target plan based on the new evaluation information.

[0042] Similarly, when there are many alternatives, and at least two of each alternative have the same score according to their respective evaluation information, the two alternatives with the same score can be used as the target solution.

[0043] The decision-making method provided in the present application obtains the evaluation information of the decision makers for each alternative plan, and then analyzes the evaluation result information of each decision maker for different alternative plans from each evaluation information, so as to determine the selection tendency of each decision maker for different alternative plans, thereby effectively helping each decision maker to determine an effective target plan.

[0044] In order to better implement the embodiment of the present application, in one embodiment of the present application, determining the evaluation result information of each evaluation information for the corresponding alternative solution includes:

[0045] Obtain evaluation information of each decision maker for each alternative plan; determine semantic information in each evaluation information; and determine evaluation result information of each evaluation information for its corresponding alternative plan based on the semantic information of each evaluation information.

[0046] The above embodiment provides a solution for determining specific evaluation result information based only on the BERT model and subsequent feature comparison. However, in actual situations, the context in the text usually has semantic information in a specific context. Therefore, in order to make the determined evaluation result information closer to the true intention of the decision maker, it is also necessary to analyze the specific semantic information.

[0047] In the embodiment of the present application, after passing through the BERT model, the word vector of the evaluation information can be obtained. At this time, the word vector can be input into the preset semantic information analysis model. For example, the E obtained after inputting the above embodiment into the BERT model s =[e i1 , e i2 ,…,e in ]∈R N*demb The word vector is input into the Bi-Lstm model to obtain h lstm =[h1,h2,…,h n ]∈R N*2d as well as Semantic features. After the semantic features are obtained, the semantic features corresponding to the evaluation information can be input into the relevant classifier to output the tendency label of the current evaluation information. For example: very inclined label, tendency label, neutral label, not inclined label and very not inclined label, etc. For example, if the label of evaluation information A1 is a very inclined label and the label of evaluation information A2 is a tendency label, it can be determined that decision maker A is more inclined to alternative 1 than alternative 2. In the embodiment of the present application, the output tendency label can be regarded as the evaluation result information.

[0048] In order to better implement the embodiment of the present application, in one embodiment of the present application, the semantic information in each evaluation information is determined, including:

[0049] Each evaluation information in each evaluation information is taken as the target evaluation information, and the syntactic features in the target evaluation information and the part-of-speech features in the target evaluation information are determined; the syntactic features and the part-of-speech features are fused to obtain fused features, and the fused features are regarded as the semantic information of the target evaluation information.

[0050] In the above embodiment, a scheme is provided for inputting the semantic information output by the Bi-Lstm model into a preset classifier to determine the evaluation result information. However, in actual situations, if the syntax and part of speech in the evaluation information are considered, the decision maker's tendency can be better obtained. Therefore, on the basis of the above embodiment, further analysis can be performed on the syntax and part of speech to determine the syntactic information and part of speech information in the evaluation information, and to perform tendency analysis as semantic information. Specifically, syntactic information can refer to a part of grammar, which discusses the relationship between predicates, modifiers and other words according to the fixed usage of the language being studied, and can refer to sentence structure. For example: sentences with exclamation marks added and sentences with progressive structures generally express stronger emotions. In addition, since dimension words usually appear in the form of nouns in the decision-making process, descriptions related to nouns (such as "nsubj") are more likely to convey emotions on the target dimension. Types of adjective-related descriptions (such as "adj-related") are often used to modify nouns to express attitudes (good or bad, etc.) towards the noun. Moreover, conjunctions indicate the same or opposite meanings of the former and the latter, which is very helpful for judging sentiment tendency (same or however, etc.), while "verb-related" indicates the action that the decision maker wants to do something (want to oppose or support, etc.). Therefore, part of speech information can also be considered on the basis of syntactic information.

[0051] Therefore, after obtaining the semantic features of the evaluation information, the specific syntactic features can be analyzed according to the semantic features. For example, the syntactic features and part-of-speech features are analyzed using a preset graph convolutional network, including: inputting the semantic features obtained in the above embodiment into the graph convolutional network to obtain syntactic features and part-of-speech features.

[0052] Alternatively, in order to better implement the embodiment of the present application, in one embodiment of the present application, determining the syntactic features in the target evaluation information includes:

[0053] Obtain an initial syntactic feature graph; obtain the syntactic distance between preset words; determine a target syntactic feature graph based on the initial syntactic feature graph and the syntactic distance between each word in the target evaluation information; determine the syntactic feature information in the target evaluation information based on the target syntactic feature graph.

[0054] In the embodiment of the present application, the initial syntactic feature graph can be a syntactic information feature graph that has been pre-set, so it can be directly obtained. At the same time, the syntactic distance between the preset words is also pre-set, and can be directly read when needed. After obtaining the initial syntactic feature information graph and the syntactic distance, the target syntactic feature graph can be obtained in the following manner, which can be specifically referred to as shown in formula ① and formula ②:

[0055]

[0056] A C =A r +A o ……②

[0057] Where n is w i and w j The distance between the syntactic dependencies between the target dimension words and other words, w i and w j Indicates the position of a word in a sentence; a v A word set representing a manually preset dimension; o is the preset initial syntactic feature map.

[0058] Afterwards, in the embodiment of the present application, according to the target syntactic feature graph, the syntactic features in the target evaluation information are determined, which can be obtained according to the following formula (1). The specific formula (1) is as follows:

[0059]

[0060] Among them, formula (1) is a loss function. Specifically, the initial input of the loss function is h0, h j l-1 and h0 belongs to is the output of the Bi-LSTM network in the above embodiment (the above embodiment uses the Bi-LSTM network output as semantic information, and this embodiment can use this as a parameter), A C is the target syntactic feature map obtained in the above description, and the normalized matrix is ​​h lstm ·ξ i =1 / (d i +1), d i is the dimension of the i-th word, W l ∈R n＊n and b l ∈R n＊2d Represents learnable weights and biases. Finally, the graph convolutional network can output syntactic features H C .

[0061] In order to better implement the embodiment of the present application, in one embodiment of the present application, determining the part-of-speech feature in the target evaluation information includes:

[0062] Determine the part of speech of each word in the target evaluation information; classify each word in the target evaluation information according to the part of speech of each word in the target evaluation information to obtain multiple word classifications; determine the part of speech feature information part of speech feature graph part of speech feature graph in the target evaluation information according to each of the word classifications.

[0063] At the same time, in addition to providing an implementation method for determining part-of-speech features through a preset graph convolutional network in the above embodiment, the embodiment of the present application also provides a method for using the Bi-LSTM network output features as part-of-speech feature calculation parameters to increase further analysis of the features, thereby determining the part-of-speech feature method, which can be specifically shown in the following formula (2):

[0064]

[0065]

[0066]

[0067] Among them, h lstm is one of the outputs of the Bi-LSTM network; It is the word classification for each word in the target evaluation information, including: H f 1. H f 2. H f 3. H f 4. H f 5. For example: H f 1 is the word with the same part of speech in the target text information. It should be noted that the words in each word category are represented by vectors; WS∈R 1*n and b S ∈R 1*1 is the weight matrix and bias; k is the number of pre-set part-of-speech feature maps. At this time, the output H F It should be noted that, in the embodiment of the present application, the part of speech in the target text information can be determined by referring to any part of speech determination method in the prior art, and the specific embodiment of the present application will not be described in detail.

[0068] After obtaining the part-of-speech feature and the syntactic feature, the two features are fused to obtain the fused feature, and then the fused feature is input into the classifier in the above embodiment to output the label of the current evaluation information. It should be noted that in the embodiment of the present application, the part-of-speech feature graph can also be determined by formula ③, and the specific formula ③ is as follows:

[0069]

[0070] Among them, k = {1, 2, 3, 4, 5} respectively corresponds to five different types of pre-set parts of speech. Therefore, k can be equal to 5 in the embodiment of the present application. i and w j Indicates the position information of the word in the sentence, and w i and w j Also a preset word.

[0071] In the embodiment of the present application, the part-of-speech feature and the syntactic feature can be fused according to formula (3), and the specific formula (3) is as follows:

[0072] H G =MLP(Concat(H C ,H F ))……(3)

[0073] The output H in formula (3) G Then, the fusion feature is input into the classifier in the above embodiment, and the corresponding label is output.

[0074] In order to better implement the embodiment of the present application, in one embodiment of the present application, based on the semantic information of each evaluation information, the evaluation result information of each evaluation information for the corresponding alternative solution is determined, including:

[0075] Features of non-target dimensions in the fused features are shielded to obtain a mask image corresponding to the processed fused features; based on the mask image, the evaluation result information of the alternative solutions corresponding to the target evaluation information is determined to obtain the evaluation result information of each evaluation information for its corresponding alternative solution.

[0076] However, the fused features include features of various dimensions, and information from some dimensions may interfere with the final label output of the classifier. Therefore, in order to improve the accuracy of the model output, the embodiment of the present application also provides a method for removing some unnecessary dimensional features from the fused features to improve the accuracy of the subsequent output labels.

[0077] Specifically, the labels corresponding to the fusion feature outputs can be removed by referring to formula (4) and formula (5). The specific formulas (4) and (5) are as follows:

[0078]

[0079]

[0080] Where a and m are the starting position of the target dimension word and the length of the target dimension in the preset dimensional space, respectively. Substituting the output in formula (4) into formula (5), we can get the mask image corresponding to the processed fusion feature. The specific mask is expressed as H G’ =[0, 0, ..., H t G’ ,…,h t+m G’ , …, 0]; For fusion features.

[0081] Afterwards, when the mask image is obtained, the specific output label can be obtained according to formula (6) and formula (7). The specific formula (6) and formula (7) are as follows:

[0082]

[0083]

[0084]

[0085] y=σ(W f z+b f )……(7)

[0086] In formula (6), γ is the attention weight for the target dimension. Then, the output z in formula (6) is substituted into formula (7) to obtain the specific label of the fusion feature, that is, the label of the evaluation information. In formula (7), h lstm is the output of the LSTM network; W f and b f are learnable weights and biases; σ is the activation function; and y is the predicted label.

[0087] It should also be noted that, in the embodiment of the present application, when training the classifier, it can also be trained by formula (8), and the specific formula (8) is as follows:

[0088]

[0089] Formula (8) is the loss function, where S is the number of sentences in the training set, and y i and Represent the actual value and the predicted value respectively. θ is all the parameters that need to be trained, and λ is the regularization coefficient. During the training process, the parameters in the classifier can be continuously adjusted until L(θ) completes convergence, and the classifier is trained.

[0090] In order to better implement the embodiment of the present application, in one embodiment of the present application, at least one target solution is determined from each alternative solution according to each evaluation result information, including:

[0091] Determine the decision factor information corresponding to each evaluation result information; and determine at least one target solution from each alternative solution based on each decision factor information.

[0092] In the above embodiment, an implementation method is provided in which each evaluation information label can be output according to a classifier. In the embodiment of the present application, a propensity score corresponding to each label can be determined according to the label, and the range of the propensity score is 0 to 1, where 0 represents a complete dislike for selecting the alternative corresponding to the evaluation information, and 1 represents 100% desire to select the alternative corresponding to the evaluation information. Among them, in the embodiment of the present application, the determination between the label and the propensity score can be determined by a preset association relationship, for example, one label corresponds to one propensity score. After the propensity score corresponding to each evaluation information is determined, a corresponding decision matrix can be formed for each decision maker. Specifically, the decision matrix is ​​as follows:

[0093]

[0094] In the decision matrix, Characterizes the propensity score of each evaluation information of the current decision maker for each alternative, for For example, decision maker A's evaluation information for alternative 1 is A1, and decision maker A's evaluation information for alternative 2 is A2. Assuming that the propensity score of A1 is 0.6 and the propensity score of A2 is 0.8, the decision matrix is ​​as follows:

[0095]

[0096] It should be noted that because the preferences obtained after comparing two identical alternatives are the same, the diagonal elements of the decision matrix are all 0.5.

[0097] Finally, according to each decision matrix, the target solution can be determined from each alternative solution. For example, according to the corresponding calculation rules, the output of the decision matrix is ​​determined, and finally a score of each decision maker for different solutions is determined, and the final target solution is determined, including: measuring the CI value of the decision matrix. If the CI value is greater than the preset threshold α, a consensus is reached and the final decision result is obtained.

[0098] In order to better implement the embodiments of the present application, in one embodiment of the present application, at least one target solution is determined from each alternative solution according to each decision matrix, including:

[0099] Clustering the evaluation information corresponding to each evaluation result information to obtain each text clustering cluster; determining part of the evaluation information in each text clustering cluster as the target evaluation information; and determining at least one target solution from each alternative solution according to the target decision factor information corresponding to each target evaluation information.

[0100] However, in actual situations, the number of decision makers may be large. If the decision matrix of each decision maker is calculated to determine the implementation method of the final target solution, more calculation data may be involved. Therefore, in order to reduce the amount of calculated data, each decision maker can be clustered according to the propensity score of each evaluation information of each decision maker to obtain a decision maker who has completed the classification. At this time, one decision maker in each classification cluster can be randomly selected as a representative decision maker. After that, the decision matrix of each representative decision maker can be calculated. That is, after the decision makers are clustered and the representative decision maker is determined, the evaluation information corresponding to the representative decision maker can be set as the target evaluation information. After that, according to the method of the above embodiment, the target decision matrix corresponding to the target evaluation information is calculated to obtain the target solution.

[0101] It should also be noted that in this process, the decision maker can manually modify the corresponding parameters in the decision matrix. For example, after obtaining the target decision matrix, the decision maker in the cluster corresponding to the target decision matrix can modify the target decision matrix. For details, please refer to Figure 3 As shown. For example: after obtaining the propensity scores of each evaluation information, each decision matrix can be generated according to the method of generating a decision matrix according to the propensity scores described in the above embodiment. At this time, each decision maker can be clustered according to each propensity score. In the embodiment of the present application, any clustering method can be used to cluster each decision maker according to the propensity score, such as K-means clustering, hierarchical clustering, density clustering, Gaussian mixture model (GMM) clustering, etc., and a specific clustering method can be used according to the actual situation, which is not limited in the embodiment of the present application. After the clustering is completed, each small group grouping C1, C2, C3, etc. can be obtained. It should be noted that in the embodiment of the present application, since each evaluation information corresponds to the identity of the decision maker. Therefore, in the process of clustering, the propensity scores corresponding to the evaluation information of the same identity need to be regarded as a whole. Furthermore, after the clustering is completed, each small group grouping includes the evaluation information of each whole. At this time, since each overall evaluation information actually corresponds to a specific decision maker, any overall evaluation information can be regarded as a representative decision maker from each small group in a random manner. Next, the decision matrix representing the decision maker can be used for subsequent calculations.

[0102] At the same time, it should be noted that the above embodiment provides a method of first determining the evaluation information corresponding to each decision maker based on the identifiers of each evaluation information, and then generating a decision matrix corresponding to each decision maker. Then, a clustering method is provided. Therefore, it can be seen that in the above embodiment, redundant decision matrices are generated. In the embodiment of the present application, it is also possible to first determine each decision maker based on the identifiers of each evaluation information. First, cluster the decision makers according to each propensity score. After clustering is completed, after selecting the representative decision maker, only the decision matrix representing the decision maker is generated.

[0103] Meanwhile, the above embodiment provides a method for calculating the CI value through the decision matrix. It should be noted that the CI value (Consistency Index) is an indicator for measuring consistency. Specifically, the method for calculating the CI value is as follows:

[0104] Sum each column of the decision matrix by row to get the sum of the weights of each column; divide each element by the sum of the weights of the corresponding column to get the normalized decision matrix; sum each row of the normalized decision matrix to get the sum of the eigenvectors of each row; for each row of the eigenvector, divide it by the sum of the eigenvectors of the corresponding row to get the normalized eigenvector; calculate the consistency index between the eigenvector and the original decision matrix, that is, CI = (maximum eigenvalue-eigenvector dimension) / (eigenvector dimension-1). Among them, the eigenvector dimension refers to the dimension (number of columns) of the decision matrix, and the maximum eigenvalue is the maximum eigenvalue obtained by eigenvalue decomposition. The CI value usually ranges from 0 to 1. The closer to 0, the better the consistency, and the closer to 1, the poorer the consistency. In order to evaluate the consistency more accurately, a random consistency index (RandomIndex, RI) can also be calculated, and the CI value is compared with the RI value to get a consistency ratio (ConsistencyRatio, CR = CI / RI). Generally, the CR value is required to be less than or equal to 0.1 to consider that the matrix has good consistency. It should be noted that this method of calculating CI and CR values ​​is applicable to a single objective or criterion in a judgment matrix. For a decision matrix with multiple objectives or a hierarchical structure, it may be necessary to apply the Analytic Hierarchy Process (AHP) or other related methods to conduct consistency assessment.

[0105] In addition, it should be noted that in the embodiment of the present application, after the representative decision maker is determined, the corresponding target decision matrix can still be corrected. For example: after the decision-making device determines the representative decision maker, it can prompt the representative decision maker whether he needs to correct his target decision matrix. If the representative decision maker does not modify the target decision matrix within the specified time, the decision-making device can decide on the target solution based on the target decision matrix according to the method described in the above embodiment; or, when the representative decision maker modifies the target decision matrix within the specified time, the decision-making device will decide on the target solution based on the new target decision matrix according to the method described in the above embodiment.

[0106] In order to better implement the decision-making method in the embodiment of the present application, a decision-making device is also provided in the embodiment of the present application on top of the decision-making method, such as Figure 4 As shown, the device 300 includes:

[0107] An acquisition module 301 is used to acquire evaluation information corresponding to each alternative solution;

[0108] A determination module 302 is used to determine evaluation result information of each evaluation information for its corresponding alternative solution;

[0109] The decision module 303 is used to determine at least one target solution from the alternative solutions according to the evaluation result information.

[0110] The decision-making device provided in the present application obtains the evaluation information of the decision maker for each alternative plan through the acquisition module 301, and then determines and analyzes the evaluation result information of each decision maker for different alternative plans according to the determination module 302 from each evaluation information, so as to determine the selection tendency of each decision maker for different alternative plans, so that the decision module 303 can effectively help each decision maker determine an effective target plan.

[0111] In some embodiments of the present application, the determination module 302 is specifically used to:

[0112] Obtain the evaluation information of each decision maker on each alternative plan;

[0113] Determine semantic information in each evaluation information;

[0114] According to the semantic information of each evaluation information, evaluation result information of each evaluation information for the corresponding alternative solution is determined.

[0115] In some embodiments of the present application, the determination module 302 is further configured to:

[0116] Taking each evaluation information in each evaluation information as target evaluation information, determining syntactic features in the target evaluation information and determining part-of-speech features in the target evaluation information;

[0117] The syntactic features and part-of-speech features are fused to obtain fused features, which are regarded as the semantic information of the target evaluation information.

[0118] In some embodiments of the present application, the determination module 302 is further configured to:

[0119] Obtain an initial syntactic feature map;

[0120] Get the syntactic distance between preset words;

[0121] Determining a target syntactic feature graph according to the initial syntactic feature graph and the syntactic distance between each word in the target evaluation information;

[0122] According to the target syntactic feature graph, syntactic feature information in the target evaluation information is determined.

[0123] In some embodiments of the present application, the determination module 302 is further configured to:

[0124] Determining the part of speech of each word in the target evaluation information;

[0125] Classifying each word in the target evaluation information according to the part of speech of each word in the target evaluation information to obtain multiple word classifications;

[0126] According to each of the word classifications, part-of-speech feature information in the target evaluation information is determined.

[0127] In some embodiments of the present application, the determination module 302 is further configured to:

[0128] Shield the features of non-target dimensions in the fused features to obtain a mask image corresponding to the processed fused features;

[0129] According to the mask image, the evaluation result information of the candidate solutions corresponding to the target evaluation information is determined to obtain the evaluation result information of each evaluation information for the corresponding candidate solutions.

[0130] In some embodiments of the present application, the decision module 303 is specifically used to:

[0131] Determine the decision matrix corresponding to each evaluation result information;

[0132] According to each decision matrix, at least one target solution is determined from each alternative solution.

[0133] In some embodiments of the present application, the decision module 303 is further configured to:

[0134] Clustering the evaluation information corresponding to each evaluation result information to obtain each text clustering cluster;

[0135] Determine part of the evaluation information in each text cluster as target evaluation information;

[0136] According to the target decision matrix corresponding to each target evaluation information, the target solution is decided from various alternative solutions.

[0137] The present application also provides a terminal device, which includes a processor, a memory, and a computer program stored in the memory and executable on the processor, and the processor executes the computer program to implement the steps in any one of the decision-making methods in the present application. The terminal device integrates any one of the decision-making methods provided in the present application. Figure 5 As shown, it shows a schematic diagram of the structure of the terminal device involved in the embodiment of the present application, specifically:

[0138] The terminal device may include one or more processing core processors 401, one or more computer-readable storage media memories 402, a power supply 403, an input unit 404 and other components. Those skilled in the art will appreciate that Figure 5 The terminal device structure shown in the figure does not constitute a limitation on the terminal device, and may include more or fewer components than shown in the figure, or combine certain components, or arrange the components differently. Among them:

[0139] The processor 401 is the control center of the terminal device, and uses various interfaces and lines to connect various parts of the entire terminal device. By running or executing software programs and / or modules stored in the memory 402, and calling data stored in the memory 402, the processor 401 executes various functions of the terminal device and processes data, thereby monitoring the terminal device as a whole. Optionally, the processor 401 may include one or more processing cores; the processor 401 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc. Preferably, the processor 401 may integrate an application processor and a modem processor, wherein the application processor mainly processes the operating system, the user interface and the application program, etc., and the modem processor mainly processes wireless communication. It is understandable that the above-mentioned modem processor may not be integrated into the processor 401.

[0140] The memory 402 can be used to store software programs and modules. The processor 401 executes various functional applications and data processing by running the software programs and modules stored in the memory 402. The memory 402 may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function (such as a sound playback function, an image playback function, etc.), etc.; the data storage area may store data created according to the use of the terminal device, etc. In addition, the memory 402 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other volatile solid-state storage devices. Accordingly, the memory 402 may also include a memory controller to provide the processor 401 with access to the memory 402.

[0141] The terminal device also includes a power supply 403 for supplying power to each component. Preferably, the power supply 403 can be logically connected to the processor 401 through a power management system, so as to manage charging, discharging, power consumption management and other functions through the power management system. The power supply 403 can also include any components such as one or more DC or AC power supplies, recharging systems, power failure detection circuits, power converters or inverters, and power status indicators.

[0142] The terminal device may further include an input unit 404, which may be used to receive input digital or character information and generate keyboard, mouse, joystick, optical or trackball signal input related to user settings and function control.

[0143] Although not shown, the terminal device may further include a display unit, etc., which will not be described in detail herein. Specifically in this embodiment, the processor 401 in the terminal device will load the executable files corresponding to the processes of one or more application programs into the memory 402 according to the following instructions, and the processor 401 will run the application programs stored in the memory 402, thereby realizing various functions, such as:

[0144] Obtain the evaluation information corresponding to each alternative plan;

[0145] Determine the evaluation result information of each evaluation information for the corresponding alternative solution;

[0146] According to the evaluation result information, at least one target solution is determined from the alternative solutions.

[0147] A person of ordinary skill in the art will appreciate that all or part of the steps in the various methods of the above embodiments may be completed by instructions, or by controlling related hardware through instructions. The instructions may be stored in a computer-readable storage medium and loaded and executed by a processor.

[0148] To this end, an embodiment of the present application provides a computer-readable storage medium, which may include: a read-only memory (ROM), a random access memory (RAM), a disk or an optical disk, etc. A computer program is stored thereon, and the computer program is loaded by a processor to execute the steps in any one of the decision-making methods provided in the embodiments of the present application. For example, the computer program may be loaded by a processor to execute the following steps:

[0149] Obtain the evaluation information corresponding to each alternative plan;

[0150] Determine the evaluation result information of each evaluation information for the corresponding alternative solution;

[0151] According to the evaluation result information, at least one target solution is determined from the alternative solutions.

[0152] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, please refer to the detailed description of other embodiments above, and will not be repeated here.

[0153] In specific implementation, the above units or structures can be implemented as independent entities, or can be arbitrarily combined to be implemented as the same or several entities. The specific implementation of the above units or structures can refer to the previous method embodiments, which will not be repeated here.

[0154] The specific implementation of the above operations can be found in the previous embodiments, which will not be described in detail here.

[0155] The above is a detailed introduction to a decision-making method and device provided in an embodiment of the present application. Specific examples are used in this article to illustrate the principles and implementation methods of the present application. The description of the above embodiments is only used to help understand the method of the present application and its core idea; at the same time, for technical personnel in this field, according to the idea of ​​the present application, there will be changes in the specific implementation method and application scope. In summary, the content of this specification should not be understood as a limitation on the present application.

Claims

1. A decision-making method, characterized in that: The method comprises: Obtain the evaluation information corresponding to each alternative plan; Determining evaluation result information of each evaluation information for the corresponding alternative solution; At least one target solution information is determined from each of the alternative solutions according to each of the evaluation result information.

2. The decision-making method according to claim 1, characterized in that: The step of determining evaluation result information of each evaluation information for the corresponding alternative solution includes: Determining semantic information in each of the evaluation information; According to the semantic information of each piece of evaluation information, evaluation result information of each piece of evaluation information for the corresponding alternative solution is determined.

3. The decision-making method according to claim 2, characterized in that: The determining of the semantic information in each of the evaluation information includes: Taking each evaluation information in the evaluation information as target evaluation information, determining syntactic feature information in the target evaluation information and determining part-of-speech feature information in the target evaluation information; The syntactic feature information and the part-of-speech feature information are subjected to feature fusion to obtain fused features, and the fused features are regarded as semantic information of the target evaluation information.

4. The decision-making method according to claim 3, characterized in that: The determining of the syntactic feature information in the target evaluation information and the determining of the part-of-speech feature information in the target evaluation information include: Obtain an initial syntactic feature map; Get the syntactic distance between preset words; Determining a target syntactic feature graph according to the initial syntactic feature graph and the syntactic distance between each word in the target evaluation information; Determining syntactic feature information in the target evaluation information according to the target syntactic feature graph; Determining the part of speech of each word in the target evaluation information; Classifying each word in the target evaluation information according to the part of speech of each word in the target evaluation information to obtain multiple word classifications; According to each of the word classifications, part-of-speech feature information in the target evaluation information is determined.

5. The decision-making method according to claim 3, characterized in that: Determining evaluation result information of each evaluation information for its corresponding alternative solution according to the semantic information of each evaluation information includes: Shielding features of non-target dimensions in the fused features to obtain a mask image corresponding to the processed fused features; The evaluation result information of the alternative solutions corresponding to the target evaluation information is determined according to the mask image, so as to obtain the evaluation result information of each evaluation information for the corresponding alternative solution.

6. The decision-making method according to claim 1, characterized in that: Determining at least one target solution from the alternative solutions according to the evaluation result information includes: Determine the decision factor information corresponding to each evaluation result information; At least one target solution is determined from the alternative solutions according to the decision factor information.

7. The decision-making method according to claim 6, characterized in that: Determining at least one target solution from the alternative solutions according to the decision factor information includes: Clustering the evaluation information corresponding to each evaluation result information to obtain each text clustering cluster; Determining part of the evaluation information in each of the text clusters as target evaluation information; At least one target solution is determined from each of the alternative solutions according to the target decision factor information corresponding to each of the target evaluation information.

8. A decision-making device, characterized in that: The device comprises: An acquisition module is used to obtain evaluation information corresponding to each alternative plan; A determination module, used to determine the evaluation result information of each evaluation information for the corresponding alternative solution; The decision module is used to determine at least one target solution from the alternative solutions according to the evaluation result information.

9. A terminal device, characterized in that: The terminal device includes a processor, a memory, and a computer program stored in the memory and executable on the processor, and the processor executes the computer program to implement the steps in the decision-making method described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, which is executed by a processor to implement the steps in the decision-making method according to any one of claims 1 to 7.

Images