A slope safety level classification method based on fuzzy language neighborhood granular concept

Abstract

This invention discloses a slope safety level classification method based on fuzzy linguistic neighborhood particle concepts, relating to the field of slope safety technology. The method includes collecting slope sample data, detecting slope condition features, constructing a linguistic concept set of slope condition features using linguistic terminology, establishing a binary relationship between slope condition features and the linguistic concept set, determining the neighborhood of the slope sample using cosine similarity, and generating an initial fuzzy linguistic neighborhood particle concept space. Based on the initial fuzzy linguistic neighborhood particle concept space, fusionable concept clusters in the subspace are labeled, and a fusion concept space is generated through weighted fusion. Based on the fusion concept space, new slope samples are classified for slope safety using minimum discriminative power. Therefore, this slope safety level classification method based on fuzzy linguistic neighborhood particle concepts is suitable for classifying the slope safety level of complex slopes, providing a reference for the flexible classification and quantification of the multi-factor coupling influence of special slopes.

Description

Technical Field

[0001] This invention relates to the field of slope safety technology, and in particular to a slope safety level classification method based on the concept of fuzzy language neighborhood particles. Background Technology

[0002] Slope safety level is an important standard for differentiating slopes based on varying geological conditions and specific project circumstances during the design and construction of slope protection projects. Currently, there are various methods for classifying slope safety levels, and the requirements for slope safety levels differ across geological conditions. Furthermore, the influencing factors are complex, making it difficult to establish a unified evaluation system that cannot adapt to the classification of slope safety levels in complex environments.

[0003] In recent years, with the development of artificial intelligence and computational intelligence technologies, some new methods such as fuzzy comprehensive evaluation, neural networks, and support vector machines have been introduced into slope stability evaluation, which has improved the accuracy of classification to some extent. However, most of these methods rely on a large amount of sample data and clear feature boundaries, and their ability to process small-sample, multi-source heterogeneous, and linguistically described slope data is still insufficient. In particular, existing methods often perform poorly in dealing with fuzzy semantic information and non-numerical empirical knowledge commonly found in complex slope engineering, which restricts their widespread application in complex environments.

[0004] Therefore, there is an urgent need for a new method for classifying slope safety levels that can effectively handle semantic ambiguity, integrate multi-source information, and have stronger interpretability and adaptability, providing new ideas for achieving more flexible, reliable and adaptive slope safety level determination. Summary of the Invention

[0005] The purpose of this invention is to provide a slope safety level classification method based on the concept of fuzzy language neighborhood particles. By combining fuzzy language modeling, neighborhood particle calculation and formal concept analysis, an intelligent classification method suitable for complex slope environments is constructed to overcome the problems existing in the prior art.

[0006] To achieve the above objectives, this invention provides a slope safety level classification method based on the concept of fuzzy language neighborhood particles, comprising the following steps: S1. Collect slope sample data, detect slope condition characteristics, and use relevant terminology. Construct a linguistic concept set of slope condition characteristics, and simultaneously construct a slope-linguistic concept binary relationship between slope condition characteristics and the linguistic concept set. ; Based on the binary relationship between slope and linguistic concepts Calculate the cosine similarity between any two sample data points, set a threshold to determine the neighborhood of the slope sample, and generate an initial fuzzy language neighborhood concept space for slope safety, containing... The number of subspaces is consistent with the total number of decision classes; S2. Based on the initial fuzzy language neighborhood concept space for slope safety, the fusionable concept clusters in the subspace based on the extensional similarity threshold are labeled, and a fusion concept space is generated by taking the union of extensions and weighting the intensions according to a preset weight formula. S3. Based on the fusion concept space, new slope samples are classified for slope safety using the minimum identification degree.

[0007] Furthermore, in S1, the binary relation The formula for calculating the membership degree of slope condition characteristics to linguistic concepts is as follows: ; In the formula, Indicates the first The characteristics of slope conditions are mapped into linguistic terms. The description, For the slope sample set, For a set of language concepts, For the corresponding slope sample, This indicates the corresponding membership degree.

[0008] Furthermore, in S1, the cosine similarity between any two sample data is calculated using the following formula: ; In the formula, Slope sample and Cosine similarity between them.

[0009] Furthermore, in S2, the concept clusters in each subspace are represented as follows: When satisfied , At that time, Marked as a fusionable concept cluster; in, Represents object learning operators; Representation of concept clusters The corresponding connotation is the fuzzy attribute.

[0010] Furthermore, S2 includes taking the union of the clusters of concepts labeled as fusionable to obtain the fusion cluster. Furthermore, the connotations corresponding to the fusionable concept clusters are assigned according to weights.

[0011] Furthermore, for a fusion cluster , The formula for weighted distribution of the corresponding connotations, representing the number of concept clusters marked as fusionable, is as follows: ; in, Represents the learning operator for objects.

[0012] Furthermore, S3 includes using triangular fuzzy membership functions to describe the new slope sample data using linguistic terms as a binary relationship between slope condition characteristics and linguistic concepts. .

[0013] Furthermore, S3 includes calculating the minimum Euclidean distance between the new slope sample and each subspace in the fused conceptual space, which serves as the discriminative power for slope safety classification. The calculation formula is as follows: ; In the formula, Indicating new slope samples In the concept space of fusion Subspace The recognizability, For subspace Central Fusion Cluster The total number, Represents the learning operator for objects.

[0014] Furthermore, the formula for calculating slope safety classification is as follows: ; In the formula, Indicating new slope samples Category code; To ensure the number of subspaces in the integrated concept space is consistent with that of the decision class.

[0015] Therefore, the present invention employs the above-mentioned slope safety level classification method based on the concept of fuzzy language neighborhood particles, which has the following technical effects: This invention comprehensively describes slope condition characteristics through the neighborhood granular concept space of slope samples, which can effectively handle semantic ambiguity in complex slope engineering, realize multi-source information fusion, enhance the interpretability of slope safety level classification, and has wide applicability.

[0016] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0017] Figure 1 This is a flowchart of a slope safety level classification method based on the concept of fuzzy language neighborhood particles; Figure 2 This is a triangular fuzzy membership function in an embodiment of a slope safety level classification method based on the concept of fuzzy language neighborhood particles. Detailed Implementation

[0018] The present invention will be explained in more detail through the following embodiments. The purpose of disclosing the present invention is to protect all changes and modifications within the scope of the present invention. The present invention is not limited to the following embodiments.

[0019] This invention is based on a quintuple. In the context of fuzzy language decision-making, the neighborhood of an object is introduced, defined as an object. In its decision-making category The most similar front A collection of objects, containing objects It itself, and the following theorem holds: Theorem 1. Let there be a quintuple. To obscure the linguistic formal context, It is a decision-based classification, given any object. neighborhood , A vague linguistic concept derived from objects.

[0020] Proof: To prove Theorem 1, it is only necessary to prove that... It is sufficient to satisfy the two properties of fuzzy language concepts, namely... and .

[0021] Based on the existing definition of conceptual cognitive learning, given a fuzzy linguistic formal context... Therefore, we can conclude that: (1) For , , ,definition The induced operator is as follows: ; ; in, Indicates all definitions in The set of all language terms on the [platform name].

[0022] (2) For any Object learning operator and attribute learning operators The definition is as follows: ; ; If the binary pair satisfy and Then it is called This refers to fuzzy language concept knowledge. (Regarding...) ,say This refers to the concept of fuzzy language particles.

[0023] From the above content, it can be seen that It is obviously satisfied. But to prove... The details are as follows: First, based on the above content, we can conclude that... ,therefore Secondly, due to If it is established, then Established.

[0024] Theorem 2. Let there be a quintuple. To obscure the linguistic formal context, It is a pair of concept learning operators, given an object neighborhood Then we have: ; .

[0025] Based on the existing definition of cognitive learning and Theorem 1, it is easy to prove that the above formula holds true, which will not be elaborated here.

[0026] Based on the above, this invention provides a slope safety level classification method based on the concept of fuzzy language neighborhood particles, including the construction of fuzzy language formal background, concept fusion method, and classification method based on minimum distinguishability, such as... Figure 1 As shown.

[0027] A. Based on existing fuzzy concept cognitive learning, a fuzzy linguistic formal background is constructed, focusing on the membership relationship between objects and fuzzy attributes, as follows: A1. Collect data from 8 slope samples and record them. This involves a non-empty finite set of objects, and the detection of relevant feature parameters for slope samples. In this embodiment, the edge condition feature set (i.e., the attribute set) is... There are six categories of slope condition characteristics, with each element representing, in order, unit weight, pore pressure, cohesion, internal friction angle, slope angle, and slope height; the slope decision feature set is as follows: , representing safety or destruction; , This represents the total number of decision categories. Specific values ​​for the slope sample data are shown in Table 1.

[0028] Table 1 Slope Sample Data

[0029] A2. Select the linguistic terminology set for slope characteristics as follows: ， It is a finite set of linguistic items consisting of an odd number of linguistic items. For positive integers; using triangular fuzzy membership functions (such as...) Figure 2As shown in the figure, a, b, and c on the horizontal axis represent the minimum, median, and maximum values ​​of the slope sample data, respectively. This data is used to model the language and incorporate the collected numerical slope sample data. ( Mapped to language terminology description In this embodiment If the value is 1, then The elements are represented in order as Small, Medium, and Large. This process yields slope sample data. Language concept set , represented as: .

[0030] A3. This embodiment uses a quintuple. As a background to fuzzy language decision-making, slope condition characteristics With language concept set Binary relations between For any Membership degree exists ,remember for ; ,in This embodiment represents the total number of decision categories. The value is 2; and Binary relations between .

[0031] In the context of fuzzy language decision-making In, any two samples and cosine similarity The calculation formula is as follows: ; Obviously, .

[0032] Since objects in similar classes are closely related, concept learning can be performed through the object's neighborhood. To this end, parameters are introduced... (i.e., the number of objects in the neighborhood), obtain the most relevant neighborhood for each object. In the decision class... ( In 2), the object Most similar front A collection of objects (containing objects) (itself), as an object neighborhood In this embodiment, a sample set with a similarity greater than 0.6 can be selected as the neighborhood. ,as follows: , , , , , , , .

[0033] Based on the aforementioned object neighborhood, a fuzzy language neighborhood granular concept space is generated. ,and Specifically, it can be expressed as follows: ; .

[0034] B. In the conceptual space, different concepts interact with each other, and there is often a lot of redundant information. To improve the accuracy of cognition, similar concepts can be found and merged in the original conceptual space. This embodiment describes a completely new fuzzy language concept from a cognitive perspective, which serves as an important variable in the subsequent concept identification process. The specific operation is as follows: B1. First of all, regarding ( Concept clusters in ) Remove redundant concepts, if the following conditions are met If it is, then it is marked as a fusionable cluster.

[0035] For example, subspace In the diagram, the fusionable clusters are labeled as follows: , , .

[0036] B2. Initial fuzzy language neighborhood concept space The concept clusters in the data are weighted and fused to generate a fused concept space. ,include: First, for any set of fusionable concept clusters ( (where is any integer), the union of the extension (i.e., the set of objects) is calculated using the following formula: ; Then, for fusionable concept clusters The connotation (i.e., fuzzy attribute) is assigned according to weight, as shown in the following formula: .

[0037] Concepts with stronger extensional generalization ability play a greater role in the fusion process. Simultaneously, the sum of the weights of all concepts is 1, meaning their overall effect is summed to 1. From a cognitive perspective, this is an effective way to describe a progressive conceptual cognition process that aligns with human cognitive patterns, and it constructs a fusion of fuzzy linguistic granular concepts. It has the following properties: (1) Arbitrary fusion of fuzzy language granular concepts It has only one parent concept, itself.

[0038] (2) Arbitrary fuzzy language granule concept There exists at least one concept of fused fuzzy language particles. extension Include .

[0039] Specifically, ; .

[0040] C. A classification method based on minimum distinguishability calculates the similarity of a given sample to each fused fuzzy language neighborhood concept subspace, thereby determining the category label.

[0041] C1. For the new slope sample data Using the triangular fuzzification formula and linguistic terminology Described as slope condition characteristics With language concept Binary relations between .

[0042] C2. Calculate new slope samples With the concept space of integration The minimum Euclidean distance between each subspace is used to calculate the discriminative power. ; in, For subspace Central Fusion Cluster The total number.

[0043] Then, based on the identification level, the new slope samples are... The categories are as follows: .

[0044] In this embodiment, the input new slope sample data is: ; for , The recognition accuracy calculation results are as follows: ; ; Therefore, new samples The decision category is Finally, the slope safety classification was completed.

[0045] Therefore, the slope safety level classification method based on the concept of fuzzy language neighborhood particles adopted in this invention can better simulate the human cognitive process of concept cognitive learning, effectively handle the uncertainty in the data, be applicable to the slope safety level classification of complex slopes, achieve more flexible, reliable and adaptive slope safety level determination, and have stronger interpretability and adaptability.

[0046] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the technical solutions of the present invention, and these modifications or equivalent substitutions cannot cause the modified technical solutions to deviate from the spirit and scope of the technical solutions of the present invention.

Claims

1. A slope safety level classification method based on the concept of fuzzy language neighborhood particles, characterized in that, Includes the following steps: S1. Collect slope sample data, detect slope condition characteristics, and construct a linguistic concept set of slope condition characteristics using linguistic terminology. Simultaneously, establish a slope-linguistic concept binary relationship between slope condition characteristics and the linguistic concept set. ; Based on the binary relationship between slope and linguistic concepts Calculate the cosine similarity between any two sample data points, set a threshold to determine the neighborhood of the slope sample, and generate an initial fuzzy language neighborhood concept space for slope safety, containing... The number of subspaces is consistent with the total number of decision classes; S2. Based on the initial fuzzy language neighborhood concept space for slope safety, the fusionable concept clusters in the subspace based on the extensional similarity threshold are labeled, and a fusion concept space is generated by taking the union of extensions and weighting the intensions according to a preset weight formula. S3. Based on the fusion concept space, new slope samples are classified for slope safety using the minimum distinguishability.

2. The slope safety level classification method based on the concept of fuzzy language neighborhood granules as described in claim 1, characterized in that, In S1, binary relation The formula for calculating the membership degree of slope condition characteristics to linguistic concepts is as follows: ； In the formula, For the corresponding slope sample, Indicates the first The characteristics of slope conditions are mapped into linguistic terms. The description, Indicates the corresponding membership degree. For the slope sample set, It is a set of language concepts.

3. The slope safety level classification method based on the concept of fuzzy language neighborhood particles according to claim 2, characterized in that, In S1, the cosine similarity between any two sample data is calculated using the following formula: ； In the formula, Slope sample and Cosine similarity between them.

4. The slope safety level classification method based on the concept of fuzzy language neighborhood granules according to claim 1, characterized in that, In S2, the concept clusters in each subspace are represented as follows: When satisfied , At that time, Marked as a fusionable concept cluster; in, Represents object learning operators; Representation of concept clusters The corresponding connotation is the fuzzy attribute.

5. The slope safety level classification method based on the concept of fuzzy language neighborhood granules according to claim 1, characterized in that, S2 involves taking the union of the clusters of concepts labeled as fusionable to obtain the fusion cluster. Furthermore, the connotations corresponding to the fusionable concept clusters are assigned according to weights.

6. The slope safety level classification method based on the concept of fuzzy language neighborhood granules according to claim 5, characterized in that, For each fusion cluster , The formula for weighted distribution of the corresponding connotations, representing the number of concept clusters marked as fusionable, is as follows: ； in, Represents the learning operator for objects.

7. The slope safety level classification method based on the concept of fuzzy language neighborhood granules according to claim 1, characterized in that, S3 includes using triangular fuzzy membership functions to describe the new slope sample data using linguistic terms as a binary relationship between slope condition characteristics and linguistic concepts. .

8. The slope safety level classification method based on the concept of fuzzy language neighborhood granules according to claim 7, characterized in that, S3 includes calculating the minimum Euclidean distance between the new slope sample and each subspace in the fused conceptual space, which serves as the distinguishability for slope safety classification. The calculation formula is as follows: ； In the formula, Indicating new slope samples In the concept space of fusion Subspace The recognizability, For subspace Central Fusion Cluster The total number, Represents the learning operator for objects.

9. The slope safety level classification method based on the concept of fuzzy language neighborhood particles according to claim 8, characterized in that, The formula for calculating slope safety classification is: ； In the formula, Indicating new slope samples Category code; To ensure the number of subspaces in the integrated concept space is consistent with that of the decision class.

Images