Water supply price reference index determination method, device and system based on large language model, and storage medium

A large language model-based method processes water supply price data by grouping, parsing, and constructing knowledge graphs to address inefficiencies in determining water prices, ensuring accurate and efficient pricing.

CN119067733AActive Publication Date: 2024-12-03BEIJING E20 ENVIRONMENT CO LTD

Patent Information

Application Number
CN202411038552.1
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-31
Publication Date
2024-12-03
Estimated Expiration
2044-07-31

AI Technical Summary

Technical Problem

Existing methods for determining water prices fail to effectively address the complexity of processing large datasets from diverse public opinions, leading to inefficiencies in determining fair pricing, and existing systems lack the capability to accurately analyze public feedback for determining water supply prices, resulting in inefficiencies in pricing determination.

Method used

A method using a large language model to process and analyze water supply price data by grouping and parsing data samples, calculating influence parameters, and constructing knowledge graphs to determine a water supply price reference index.

Benefits of technology

The method provides a rapid and accurate determination of water supply prices by effectively processing large datasets, improving the representativeness and authenticity of pricing decisions.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119067733A_ABST
    Figure CN119067733A_ABST
Patent Text Reader

Abstract

The invention relates to a water supply price reference index determination method, device and system based on a large language model, and a storage medium. The method comprises the steps of obtaining a first water supply price language data sample; grouping the individuals, deleting the first water supply price language data samples corresponding to invalid individual groups, and obtaining second water supply price language data samples; analyzing each second water supply price language data sample to obtain an analysis tree; grouping the second water supply price language data samples according to the analytic tree; calculating grouping weights and sorting the grouping weights; and calculating influence parameters of the analytic tree, correcting the grouping weight by using the influence parameters, and taking the corrected grouping weight as a water supply price reference index. According to the water supply price reference index determination method, device and system based on the large language model and the storage medium disclosed by the invention, the arrangement and analysis of mass collected data are realized through cleaning and normalization processing of the collected data, and the result is obtained.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of data processing technology, and in particular to a method, device, system and storage medium for determining a water supply price reference index based on a large language model. Background Art

[0002] The basic principles for determining water supply prices are cost coverage, reasonable returns, water conservation and fair burden. When local governments actually set local water supply prices, they use water supply costs as the basis, and also need to consider multiple influencing factors such as local water resources, economic development, user affordability and public feedback to ultimately determine the water supply price.

[0003] Objective influencing factors such as water resource conditions, economic development and user affordability can be obtained by organizing, calculating and analyzing data from official public channels. However, public feedback is generated by different individuals and is difficult to obtain. A very large sample size is required to obtain more accurate results. After obtaining the samples, analysis is still required to obtain results.

[0004] There are certain limitations in manual analysis, which are mainly reflected in the slow processing speed and high processing cost of massive data, and the collected data is mixed with a certain amount of useless information. This means that this analysis method can only be used in a small range, and there are certain deficiencies in representativeness and authenticity.

[0005] It is also necessary to explore a technical solution that can quickly process massive samples of public feedback. Summary of the invention

[0006] The present application provides a method, device, system and storage medium for determining a water supply price reference index based on a large language model, which can organize, analyze and obtain results of massive collected data by cleaning and normalizing the collected data.

[0007] The above-mentioned purpose of the present application is achieved through the following technical solutions:

[0008] In a first aspect, the present application provides a method for determining a water supply price reference index based on a large language model, comprising:

[0009] Obtaining a first water supply price language data sample, the number of the first water supply price language data samples is multiple, and each first water supply price language data sample comes from an independent individual;

[0010] Grouping the individuals to obtain a valid individual group and an invalid individual group, deleting the first water supply price language data sample corresponding to the invalid individual group, and obtaining a second water supply price language data sample;

[0011] Parse each second water supply price language data sample to obtain a parse tree;

[0012] Grouping the second water supply price language data samples according to the parse tree, the groups including a first sample group, a second sample group and a third sample group;

[0013] Calculate and sort the grouping weights, where the grouping weight is the ratio of the number of second water supply price language data samples in the group to the number of all second water supply price language data samples;

[0014] Calculate the influence parameters of the parse tree and use the influence parameters to correct the grouping weights;

[0015] The corrected grouping weights are used as the reference index for water supply prices.

[0016] In a possible implementation manner of the first aspect, grouping individuals includes:

[0017] Determine the first water supply price language data sample corresponding to the individual, recorded as the analysis sample;

[0018] Parsing the analysis sample to obtain the subjects included in the analysis sample;

[0019] Confirm the association relationship grid between the subjects and evaluate the association relationship grid to obtain the association relationship grid value;

[0020] Use the association grid values ​​to group individuals.

[0021] In a possible implementation manner of the first aspect, evaluating the association relationship grid to obtain the association relationship grid value includes:

[0022] Determine the nodes of the association relationship grid and the relationship between the nodes;

[0023] Use nodes to generate a topological relationship graph in two-dimensional space or three-dimensional space. The distance between two directly connected nodes is fixed, and the directions of two directly connected nodes are assigned according to attributes.

[0024] Calculate the surface area of ​​the topological relationship graph and obtain the surface area value;

[0025] The surface area value of the topological relationship graph is used as the value of the association relationship grid.

[0026] In a possible implementation manner of the first aspect, the second water supply price language data sample is parsed to obtain a parsing tree including:

[0027] Split the second water supply price language data sample into sentences;

[0028] Divide the sentence to obtain the subject and relation words included in the sentence;

[0029] Build a knowledge graph using subject and relation words and output the knowledge graph as a parse tree.

[0030] In a possible implementation of the first aspect, constructing a knowledge graph using subjects and relational words includes:

[0031] The subjects are divided into levels according to the relational terms, and each level includes at least one subject;

[0032] The connection relationship between two subjects is constructed based on the relational terms, and the two subjects belong to two adjacent levels respectively;

[0033] When a subject has a connection relationship with multiple subjects at the next level, the same subjects in the multiple subjects at the next level are merged;

[0034] Directions are assigned to subjects according to their type. Subjects of the same type are assigned the same direction, and one direction is assigned to subjects of only one type.

[0035] Use Direction to adjust the orientation of the connection between the main body and the first body and the main body.

[0036] In a possible implementation manner of the first aspect, the method further includes:

[0037] Determine the branches of the knowledge graph, which include main branches and secondary branches;

[0038] Calculate the center segment of each main branch;

[0039] Drive the main branches of the knowledge graph to rotate so that the minimum distance between any two main branches belonging to the same knowledge graph is maximized;

[0040] Among them, in all knowledge graphs, subjects of the same type are given the same direction. When a subject in a knowledge graph is adjusted, subjects of the same type in other knowledge graphs are adjusted synchronously.

[0041] In a possible implementation manner of the first aspect, calculating the influencing parameter of the parse tree includes:

[0042] Construct a spatial model based on the parse tree and calculate the volume of the spatial model;

[0043] Determine the number of main branches of the parse tree;

[0044] The volume of the spatial model and the number of main branches of the tree are weighted to obtain the influence parameters of the parse tree.

[0045] In a second aspect, the present application provides a device for determining a water supply price reference index based on a large language model, comprising:

[0046] A data acquisition unit, used for acquiring a first water supply price language data sample, the number of the first water supply price language data sample is multiple, and each first water supply price language data sample comes from an independent individual;

[0047] A first grouping unit is used to group individuals to obtain a valid individual group and an invalid individual group, delete the first water supply price language data sample corresponding to the invalid individual group, and obtain a second water supply price language data sample;

[0048] A first data processing unit is used to parse each second water supply price language data sample to obtain a parse tree;

[0049] A second grouping unit is used to group the second water supply price language data samples according to the parse tree, and the groups include a first sample group, a second sample group and a third sample group;

[0050] A first calculation unit is used to calculate and sort the grouping weight, where the grouping weight is the ratio of the number of second water supply price language data samples in the group to the number of all second water supply price language data samples;

[0051] A second calculation unit is used to calculate the influence parameters of the parse tree and use the influence parameters to correct the grouping weights;

[0052] The first result output unit is used to use the corrected grouping weight as a water supply price reference index.

[0053] In a third aspect, the present application provides a water supply price reference index determination system based on a large language model, the system comprising:

[0054] one or more memories for storing instructions; and

[0055] One or more processors, used to call and run the instructions from the memory to execute the method as described in the first aspect and any possible implementation of the first aspect.

[0056] In a fourth aspect, the present application provides a computer-readable storage medium, the computer-readable storage medium comprising:

[0057] Program, when the program is executed by a processor, the method described in the first aspect and any possible implementation of the first aspect is executed.

[0058] In a fifth aspect, the present application provides a computer program product, comprising program instructions. When the program instructions are executed by a computing device, the method described in the first aspect and any possible implementation of the first aspect is executed.

[0059] In a sixth aspect, the present application provides a chip system, which includes a processor for implementing the functions involved in the above aspects, for example, generating, receiving, sending, or processing the data and / or information involved in the above methods.

[0060] The chip system may be composed of chips, or may include chips and other discrete devices.

[0061] In a possible design, the chip system also includes a memory, which is used to store necessary program instructions and data. The processor and the memory can be decoupled and respectively set on different devices, connected by wired or wireless means, or the processor and the memory can also be coupled on the same device. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] Figure 1 This is a schematic flowchart of the steps of a method for determining a water supply price reference index provided in this application.

[0063] Figure 2 This is a schematic diagram of obtaining a surface area value of a topological relationship diagram provided by the present application.

[0064] Figure 3 This is a schematic diagram of a processing process of a first water supply price language data sample provided by the present application.

[0065] Figure 4 It is a schematic diagram of the cross-sectional area of ​​a branch in the extension direction provided by the present application, and the cross-sectional area is smaller than the reference value.

[0066] Figure 5 It is a schematic diagram of the cross-sectional area of ​​a branch in the extension direction provided by the present application, and the cross-sectional area is larger than the reference value. DETAILED DESCRIPTION

[0067] The technical solution in this application is further described in detail below in conjunction with the accompanying drawings.

[0068] See also Figure 1 In some examples, the specific steps of the method for determining the water supply price reference index based on the large language model disclosed in this application are as follows:

[0069] S101, obtaining a first water supply price language data sample, the number of the first water supply price language data samples is multiple, and each first water supply price language data sample comes from an independent individual;

[0070] S102, grouping individuals to obtain a valid individual group and an invalid individual group, deleting the first water supply price language data sample corresponding to the invalid individual group, and obtaining a second water supply price language data sample;

[0071] S103, parsing each second water supply price language data sample to obtain a parsing tree;

[0072] S104, grouping the second water supply price language data samples according to the parse tree, the groups including a first sample group, a second sample group and a third sample group;

[0073] S105, calculating and sorting the grouping weights, where the grouping weights are the ratio of the number of second water supply price language data samples in the group to the number of all second water supply price language data samples;

[0074] S106, calculating the influence parameters of the parse tree and correcting the grouping weights using the influence parameters;

[0075] S107, using the corrected grouping weight as a water supply price reference index.

[0076] The content in step S101 and step S102, wherein the first water supply price language data sample refers to the different water price-related data content collected, needs to be grouped first. There are two groups here, namely the valid individual group and the invalid individual group. Then the first water supply price language data sample corresponding to the invalid individual group is deleted to obtain the second water supply price language data sample.

[0077] The data content can be public opinion, articles, analysis reports, etc. In addition, the time period needs to be limited for data collection to avoid the reverse impact caused by cross-time periods, because the prices in different time periods and the individual's perception of prices are different.

[0078] For the first water supply price language data sample corresponding to the invalid individual group, generally speaking, only results such as price increase, price reduction and some invalid content are released. For the invalid content, blocked words can be used for filtering.

[0079] In some examples, individuals are grouped in the following ways:

[0080] S201, determining a first water supply price language data sample corresponding to an individual, recorded as an analysis sample;

[0081] S202, parsing the analysis sample to obtain the subject included in the analysis sample;

[0082] S203, confirming the association relationship grids between the subjects and evaluating the association relationship grids to obtain the association relationship grid values;

[0083] S204, grouping individuals using the association relationship grid values.

[0084] In step S201 to step S204, an association relationship grid is constructed according to the association relationship between the subjects in the first water supply price language data sample, such as Figure 2 As shown, the association grid values ​​are then calculated and finally the individuals are grouped using the association grid values.

[0085] This is a quantifiable grouping method, and the specific explanation of the relevant content is as follows:

[0086] Obtaining the subject included in the analysis sample refers to determining the content that the analysis sample wants to express. This involves natural semantic processing. A suitable processing method is to first divide the analysis sample into units, where the length of a unit is one or two words, and then try to understand the meaning of the unit, and then understand the meaning of a sentence, until the meaning of the analysis sample is understood.

[0087] The role of unit division is to determine the composition of the analysis sample. The analysis sample consists of attributive, subject, adverbial, predicate, object and complement. By dividing the analysis sample into units, the attributive, subject, adverbial, predicate, object and complement can be determined. The main body included in the analysis sample refers to the subject and the object.

[0088] After determining the subjects included in the analysis sample and the association relationships between the subjects, the association relationship grid can be obtained. Then the association relationship grid is evaluated to obtain the association relationship grid value. The specific process is as follows:

[0089] S301, determining nodes of an association relationship grid and the relationship between the nodes;

[0090] S302, using nodes to generate a topological relationship graph in a two-dimensional space or a three-dimensional space, wherein the distance between two directly connected nodes is fixed, and the two directly connected nodes are assigned directions according to attributes;

[0091] S303, calculating the surface area of ​​the topological relationship graph to obtain a surface area value;

[0092] S304, using the surface area value of the topological relationship graph as the association relationship grid value.

[0093] The node in step S301 refers to the subject mentioned in step S201 to step S204. The way to use the node to generate a topological relationship diagram in a two-dimensional space or a three-dimensional space is to assign coordinates to the node, and then determine the connection relationship between the nodes based on the belonging relationship between the nodes.

[0094] For the same and similar subjects, they need to be merged.

[0095] The connection relationship here refers to a direct connection relationship. The distance between two nodes in a direct connection relationship is fixed, for example, both are a unit length. The two nodes in a direct connection relationship are allocated directions according to attributes.

[0096] According to the specific method of attribute allocation direction, for example, the first node and the second node are used to distinguish here. The first node and multiple second nodes have a connection relationship. The meanings of the multiple second nodes are different. The meaning of the second node here is the attribute between the two nodes.

[0097] A direction is assigned to each second node. When the distance between the two nodes is fixed, the position of the second node can be determined. This is because the position of the second node can be determined through direction and distance under the premise that the position of the first node is determined.

[0098] At this time, a topological relationship diagram will be obtained in two-dimensional space or three-dimensional space, with the surface area value of the topological relationship diagram ( Figure 2 Based on this, the method of grouping individuals using the association grid value is to specify a value, and individuals greater than the value are classified into the valid individual group, and individuals less than the value are classified into the invalid individual group.

[0099] It should be understood that for the first water supply price language data sample, the effective content cannot be determined simply by the number of words, because this will cause the second water supply price language data sample to be inaccurate. For example, a first water supply price language data sample has a large number of words, but it is only an expression of personal emotions or a statement of a specific content, which makes the first water supply price language data sample have obvious tendencies and limitations.

[0100] In the present application, this first water supply price language data sample needs to be discarded, and what needs to be retained is the content stated from multiple aspects, because the water supply price is affected by multiple factors and needs to be explained from multiple aspects.

[0101] After obtaining the second water supply price language data sample, step S103 is executed. In this step, each second water supply price language data sample is parsed to obtain a parsing tree, and then the second water supply price language data sample is grouped according to the parsing tree. The groups include a first sample group, a second sample group and a third sample group. Figure 3 , which is the content in step S104.

[0102] One function of the parse tree is to group, and then calculate each group to get the final water supply price reference index. The parse tree determines the grouping basis by including keywords (water price increase, price reduction, maintenance or related).

[0103] In step S105, it is necessary to calculate the grouping weight and sort, the grouping weight is the ratio of the number of second water supply price language data samples in the group to the number of all second water supply price language data samples. The role of the grouping weight is to make the sorting not only affected by quantity and quality.

[0104] Because if we only consider the ranking from the perspective of quantity, the result will generally be no price increase or even a price reduction, because the vast majority of individuals express themselves based on their own positions.

[0105] Obviously, the result obtained at this time is not accurate, so it needs to be corrected.

[0106] In step S106, the influencing parameters of the parse tree are calculated and the grouping weights are corrected using the influencing parameters. Finally, in step S107, the corrected grouping weights are used as the water supply price reference index.

[0107] In some examples, the second water supply price language data sample is parsed to obtain a parsing tree in the following manner:

[0108] Split the second water supply price language data sample into sentences;

[0109] Divide the sentence to obtain the subject and relation words included in the sentence;

[0110] Build a knowledge graph using subject and relation words and output the knowledge graph as a parse tree.

[0111] At this time, it is still necessary to obtain the subject and relationship words, and then use the subject and relationship words to build a knowledge graph. The content of obtaining the subject and relationship words has been stated in the previous content and will not be repeated here.

[0112] The specific way to construct a knowledge graph using subject and relation words is:

[0113] The subjects are divided into levels according to the relational terms, and each level includes at least one subject;

[0114] The connection relationship between two subjects is constructed based on the relational terms, and the two subjects belong to two adjacent levels respectively;

[0115] When a subject has a connection relationship with multiple subjects at the next level, the same subjects in the multiple subjects at the next level are merged;

[0116] Directions are assigned to subjects according to their type. Subjects of the same type are assigned the same direction, and one direction is assigned to subjects of only one type.

[0117] Use Direction to adjust the orientation of the connection between the main body and the first body and the main body.

[0118] The purpose of using direction here is to transform the knowledge graph from two-dimensional space to three-dimensional space. The core reason is that two-dimensional space is limited in expressing relationships. For example, when it is required to assign directions to multiple subjects at the next level, it is necessary to consider the smaller angle caused by the number of subjects. In this case, the knowledge graph obtained in two-dimensional space will be crowded in the expansion direction.

[0119] Furthermore, the following contents are added:

[0120] Determine the branches of the knowledge graph, which include main branches and secondary branches;

[0121] Calculate the center segment of each main branch;

[0122] Drive the main branches of the knowledge graph to rotate so that the minimum distance between any two main branches belonging to the same knowledge graph is maximized;

[0123] Among them, in all knowledge graphs, subjects of the same type are given the same direction. When a subject in a knowledge graph is adjusted, subjects of the same type in other knowledge graphs are adjusted synchronously.

[0124] It should be noted that after obtaining the knowledge graph, it is also necessary to normalize the knowledge graph so that it can be compared and calculated. At this time, it is necessary to first determine the branches of the knowledge graph, which include main branches and secondary branches. The specific determination method is based on the cross-sectional area of ​​the branch in the extension direction.

[0125] For cross-sectional areas, see Figure 4 and Figure 5 The dotted line in the figure sets a reference value. Branches that are larger than the reference value are main branches, and branches that are smaller than the reference value are secondary branches.

[0126] Then the center line segment of each main branch is calculated, and then the main branch of the knowledge graph is driven to rotate, so that the minimum distance between any two main branches belonging to the same knowledge graph is maximized, so as to show the maximum difference of the knowledge graph.

[0127] In the above method, it is also necessary to assign the same direction to subjects of the same type in all knowledge graphs, and the purpose is also to achieve normalization of the knowledge graph.

[0128] In addition, when adjusting a subject in a knowledge graph, the same type of subjects in other knowledge graphs are adjusted synchronously, that is, for a subject of a type, its direction is the same in any knowledge graph. The minimum distance between any two main branches is the largest.

[0129] In some examples, the way to calculate the impact parameters of the parse tree is as follows:

[0130] Construct a spatial model based on the parse tree and calculate the volume of the spatial model;

[0131] Determine the number of main branches of the parse tree;

[0132] The volume of the spatial model and the number of main branches of the tree are weighted to obtain the influence parameters of the parse tree.

[0133] In this method, two parameters are introduced, namely, the volume of the spatial model and the number of main branches. The larger the volume of the spatial model, the more content the knowledge graph includes, and the larger the number of main branches, the more types of content the knowledge graph includes.

[0134] The weighted calculation method is to calculate by fixed value, the formula is as follows:

[0135] The influencing parameter of the parse tree = space model volume * N1 + number of main branches * N2, the sum of N1 and N2 is 1, both of which are preset values.

[0136] After obtaining the influencing parameters of the parse tree, the influencing parameters are used to correct the grouping weights. The specific method is to replace the number of samples of the second water supply price language data with the accumulation of the influencing parameters of the parse tree.

[0137] For example, the number of samples of the second water supply price language data is three, and the sample number value at this time is 3. The influencing parameters of the obtained parse tree are: 1, 3, and 7 respectively. At this time, the cumulative influencing parameters of the parse tree of the second water supply price language data sample are 11.

[0138] There are multiple second water supply price language data samples, and each second water supply price language data sample corresponds to an influencing parameter of a parse tree.

[0139] The influencing parameter of the parse tree can also be interpreted as the weight of the second water supply price language data sample.

[0140] Through the above method, a ratio of K1:K2:K3 can be obtained, where K1 corresponds to price increase (the accumulation of influencing parameters of the parse tree of the second water supply price language data sample in the first sample group), K2 corresponds to price reduction (the accumulation of influencing parameters of the parse tree of the second water supply price language data sample in the second sample group), and K3 corresponds to unchanged (the accumulation of influencing parameters of the parse tree of the second water supply price language data sample in the third sample group).

[0141] Through K1:K2:K3, we can further determine the corresponding proportions of price increases, price reductions and unchanged prices, and then choose the appropriate price adjustment plan based on this.

[0142] For example, combined with the content recorded in the previous article, K1:K2:K3 reflects the evaluation of water supply prices over a period of time. When determining the water supply price in the next stage, K1:K2:K3 is used as a reference factor. For example, if the proportion of people who agree to increase the price is high, you can consider raising the water supply price appropriately. If the proportion of people who disagree with the price increase is high, you need to choose not to adjust the water supply price or to raise the water supply price slightly.

[0143] It should also be noted that the water supply price reference index provided in this application is only used as one of the influencing factors in determining the water supply price. It is a reference and cannot directly determine the water supply price. Because when determining the water supply price, the more key influencing factors provided, the more reasonable the price will be.

[0144] The present application also provides a device for determining a water supply price reference index based on a large language model, comprising:

[0145] A data acquisition unit, used for acquiring a first water supply price language data sample, the number of the first water supply price language data sample is multiple, and each first water supply price language data sample comes from an independent individual;

[0146] A first grouping unit is used to group individuals to obtain a valid individual group and an invalid individual group, delete the first water supply price language data sample corresponding to the invalid individual group, and obtain a second water supply price language data sample;

[0147] A first data processing unit is used to parse each second water supply price language data sample to obtain a parse tree;

[0148] A second grouping unit is used to group the second water supply price language data samples according to the parse tree, and the groups include a first sample group, a second sample group and a third sample group;

[0149] A first calculation unit is used to calculate and sort the grouping weight, where the grouping weight is the ratio of the number of second water supply price language data samples in the group to the number of all second water supply price language data samples;

[0150] A second calculation unit is used to calculate the influence parameters of the parse tree and use the influence parameters to correct the grouping weights;

[0151] The first result output unit is used to use the corrected grouping weight as a water supply price reference index.

[0152] Furthermore, it also includes:

[0153] A first determining unit, used to determine a first water supply price language data sample corresponding to an individual, recorded as an analysis sample;

[0154] A second data processing unit is used to parse the analysis sample to obtain the subject included in the analysis sample;

[0155] The second determining unit is used to confirm the association relationship grid between the subjects and evaluate the association relationship grid to obtain the association relationship grid value;

[0156] The grouping unit is used to group individuals using the associated relationship grid values.

[0157] Furthermore, it also includes:

[0158] A third determining unit, used to determine the nodes of the association relationship grid and the belonging relationship between the nodes;

[0159] A generating unit, used for generating a topological relationship graph in a two-dimensional space or a three-dimensional space using nodes, wherein a distance between two directly connected nodes is fixed, and directions of two directly connected nodes are allocated according to attributes;

[0160] The third calculation unit is used to calculate the surface area of ​​the topological relationship graph to obtain a surface area value;

[0161] The second result output unit is used to use the surface area value of the topological relationship graph as the association relationship grid value.

[0162] Furthermore, it also includes:

[0163] A sentence splitting unit, used for splitting the second water supply price language data sample into sentences;

[0164] A sentence division unit, used for dividing the sentence to obtain the subject and relation words included in the sentence;

[0165] The third result output unit is used to construct a knowledge graph using subject and relation words and output the knowledge graph as a parse tree.

[0166] Furthermore, it also includes:

[0167] A level division unit, used to divide the subjects into levels according to the relational words, each level including at least one subject;

[0168] A relationship construction unit is used to construct a connection relationship between two subjects according to a relationship word, and the two subjects belong to two adjacent levels respectively;

[0169] A merging processing unit, configured to merge the same subjects among the multiple subjects at the next level when one subject has a connection relationship with the multiple subjects at the next level;

[0170] A direction assigning unit, used for assigning a direction to a subject according to the subject type. Subjects of the same type are assigned the same direction, and one direction is only assigned to subjects of one type.

[0171] The direction adjustment unit is used to adjust the connection orientation direction between the main body and the first main body and the main body.

[0172] Furthermore, it also includes:

[0173] A first processing unit is used to determine branches of the knowledge graph, where the branches include main branches and secondary branches;

[0174] A second processing unit is used to calculate the center line segment of each main branch;

[0175] An adjustment unit, used to drive the main branches of the knowledge graph to rotate so that the minimum distance between any two main branches belonging to the same knowledge graph is maximized;

[0176] Among them, in all knowledge graphs, subjects of the same type are given the same direction. When a subject in a knowledge graph is adjusted, subjects of the same type in other knowledge graphs are adjusted synchronously.

[0177] Furthermore, it also includes:

[0178] A volume calculation unit, used for constructing a spatial model based on a parse tree and calculating the volume of the spatial model;

[0179] A quantity determination unit, used to determine the number of main branches of the parse tree;

[0180] The weighted calculation unit is used to perform weighted calculation on the volume of the spatial model and the number of main branches of the tree to obtain the influencing parameters of the parse tree.

[0181] In one example, the unit in any of the above devices can be one or more integrated circuits configured to implement the above methods, such as: one or more application specific integrated circuits (ASICs), or, one or more digital signal processors (DSPs), or, one or more field programmable gate arrays (FPGAs), or a combination of at least two of these integrated circuit forms.

[0182] For another example, when the units in the device can be implemented in the form of a processing element scheduling program, the processing element can be a general-purpose processor, such as a central processing unit (CPU) or other processor that can call a program. For another example, these units can be integrated together and implemented in the form of a system-on-a-chip (SOC).

[0183] Various objects such as various messages / information / equipment / network elements / systems / devices / actions / operations / processes / concepts that may appear in this application are named. It can be understood that these specific names do not constitute a limitation on the relevant objects. The names assigned may change with factors such as scenarios, contexts or usage habits. The understanding of the technical meaning of the technical terms in this application should be mainly determined from the functions and technical effects embodied / executed in the technical scheme.

[0184] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0185] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0186] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0187] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0188] It should also be understood that in various embodiments of the present application, the first, second, etc. are only used to indicate that multiple objects are different. For example, the first time window and the second time window are only used to indicate different time windows. They should not have any impact on the time window itself, and the first, second, etc. mentioned above should not impose any limitations on the embodiments of the present application.

[0189] It should also be understood that in the various embodiments of the present application, unless otherwise specified or there is a logical conflict, the terms and / or descriptions between different embodiments are consistent and can be referenced to each other, and the technical features in different embodiments can be combined to form new embodiments according to their internal logical relationships.

[0190] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a computer-readable storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods described in each embodiment of the present application. The aforementioned computer-readable storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0191] The present application also provides a water supply price reference index determination system based on a large language model, the system comprising:

[0192] one or more memories for storing instructions; and

[0193] One or more processors are used to call and run the instructions from the memory to execute the method as described above.

[0194] The present application also provides a computer program product, which includes instructions. When the instructions are executed, the terminal device and the network device perform operations of the terminal device and the network device corresponding to the above method.

[0195] The present application also provides a chip system, which includes a processor for implementing the functions involved in the above content, such as generating, receiving, sending, or processing the data and / or information involved in the above method.

[0196] The chip system may be composed of chips, or may include chips and other discrete devices.

[0197] The processor mentioned in any of the above places can be a CPU, a microprocessor, an ASIC, or one or more integrated circuits for executing programs for controlling the above-mentioned feedback information transmission method.

[0198] In a possible design, the chip system also includes a memory, which is used to store necessary program instructions and data. The processor and the memory can be decoupled and respectively set on different devices, connected by wire or wireless means to support the chip system to implement various functions in the above embodiments. Alternatively, the processor and the memory can also be coupled on the same device.

[0199] Optionally, the computer instructions are stored in a memory.

[0200] Optionally, the memory is a storage unit within the chip, such as a register, a cache, etc. The memory can also be a storage unit within the terminal located outside the chip, such as a ROM or other types of static storage devices that can store static information and instructions, RAM, etc.

[0201] It can be understood that the memory in the present application can be a volatile memory or a non-volatile memory, or can include both volatile and non-volatile memories.

[0202] The non-volatile memory may be a ROM, a programmable ROM (PROM), an erasable PROM (EPROM), an electrically erasable PROM (EEPROM), or a flash memory.

[0203] The volatile memory may be a RAM, which is used as an external cache. There are many different types of RAM, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate synchronous dynamic random access memory (DDR SDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous link dynamic random access memory (SLDRAM) and direct memory bus random access memory.

[0204] The embodiments of this specific implementation method are all preferred embodiments of the present application, and are not intended to limit the protection scope of the present application. Therefore, all equivalent changes made based on the structure, shape, and principle of the present application should be included in the protection scope of the present application.

Claims

1. A method for determining a water supply price reference index based on a large language model, characterized in that: include: Obtaining a first water supply price language data sample, the number of the first water supply price language data samples is multiple, and each first water supply price language data sample comes from an independent individual; Grouping the individuals to obtain a valid individual group and an invalid individual group, deleting the first water supply price language data sample corresponding to the invalid individual group, and obtaining a second water supply price language data sample; Parse each second water supply price language data sample to obtain a parse tree; Grouping the second water supply price language data samples according to the parse tree, the groups including a first sample group, a second sample group and a third sample group; Calculate and sort the grouping weights, where the grouping weight is the ratio of the number of second water supply price language data samples in the group to the number of all second water supply price language data samples; Calculate the influence parameters of the parse tree and use the influence parameters to correct the grouping weights; The corrected grouping weights are used as the reference index for water supply prices.

2. The method for determining a water supply price reference index based on a large language model according to claim 1, characterized in that: Grouping individuals includes: Determine the first water supply price language data sample corresponding to the individual, recorded as the analysis sample; Parsing the analysis sample to obtain the subjects included in the analysis sample; Confirm the association relationship grid between the subjects and evaluate the association relationship grid to obtain the association relationship grid value; Use the association grid values ​​to group individuals.

3. The method for determining a water supply price reference index based on a large language model according to claim 2, characterized in that: Evaluate the association relationship grid and obtain the association relationship grid values ​​including: Determine the nodes of the association relationship grid and the relationship between the nodes; Use nodes to generate a topological relationship graph in two-dimensional space or three-dimensional space. The distance between two directly connected nodes is fixed, and the directions of two directly connected nodes are assigned according to attributes. Calculate the surface area of ​​the topological relationship graph and obtain the surface area value; The surface area value of the topological relationship graph is used as the value of the association relationship grid.

4. The method for determining a water supply price reference index based on a large language model according to any one of claims 1 to 3, characterized in that: The second water supply price language data sample is parsed to obtain a parsing tree including: Split the second water supply price language data sample into sentences; Divide the sentence to obtain the subject and relation words included in the sentence; Build a knowledge graph using subject and relation words and output the knowledge graph as a parse tree.

5. The method for determining a water supply price reference index based on a large language model according to claim 4, characterized in that: Building a knowledge graph using subject and relation words includes: The subjects are divided into levels according to the relational terms, and each level includes at least one subject; The connection relationship between two subjects is constructed based on the relational terms, and the two subjects belong to two adjacent levels respectively; When a subject has a connection relationship with multiple subjects at the next level, the same subjects in the multiple subjects at the next level are merged; Directions are assigned to subjects according to their type. Subjects of the same type are assigned the same direction, and one direction is assigned to subjects of only one type. Use Direction to adjust the orientation of the connection between the main body and the first body and the main body.

6. The method for determining a water supply price reference index based on a large language model according to claim 5, characterized in that: Also includes: Determine the branches of the knowledge graph, which include main branches and secondary branches; Calculate the center segment of each main branch; Drive the main branches of the knowledge graph to rotate so that the minimum distance between any two main branches belonging to the same knowledge graph is maximized; Among them, in all knowledge graphs, subjects of the same type are given the same direction. When a subject in a knowledge graph is adjusted, subjects of the same type in other knowledge graphs are adjusted synchronously.

7. The method for determining a water supply price reference index based on a large language model according to claim 6, characterized in that: The parameters that influence the calculation of the parse tree include: Construct a spatial model based on the parse tree and calculate the volume of the spatial model; Determine the number of main branches of the parse tree; The volume of the spatial model and the number of main branches of the tree are weighted to obtain the influence parameters of the parse tree.

8. A device for determining a water supply price reference index based on a large language model, characterized in that: include: A data acquisition unit, used for acquiring a first water supply price language data sample, the number of the first water supply price language data sample is multiple, and each first water supply price language data sample comes from an independent individual; A first grouping unit is used to group individuals to obtain a valid individual group and an invalid individual group, delete the first water supply price language data sample corresponding to the invalid individual group, and obtain a second water supply price language data sample; A first data processing unit is used to parse each second water supply price language data sample to obtain a parse tree; A second grouping unit is used to group the second water supply price language data samples according to the parse tree, and the groups include a first sample group, a second sample group and a third sample group; A first calculation unit is used to calculate and sort the grouping weight, where the grouping weight is the ratio of the number of second water supply price language data samples in the group to the number of all second water supply price language data samples; A second calculation unit is used to calculate the influence parameters of the parse tree and use the influence parameters to correct the grouping weights; The first result output unit is used to use the corrected grouping weight as a water supply price reference index.

9. A water supply price reference index determination system based on a large language model, characterized in that: The system comprises: one or more memories for storing instructions; and One or more processors, configured to call and execute the instructions from the memory to perform the method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer readable storage medium comprises: The program, when the program is executed by a processor, the method according to any one of claims 1 to 7 is executed.

Citation Information

Patent Citations

  • Data pricing method and device, electronic equipment and storage medium

    CN114255067A

  • Method and device for evaluating intention classification model, equipment and medium

    CN115048519A

  • Log analysis method and device and electronic equipment

    CN115994073A

  • Data processing method, device and system based on knowledge graph and storage medium

    CN117807252A

  • Water demand prediction method and system

    JP2020112847A

Cited By

  • Water price reference index determination method and device based on multi-element mutual feedback optimization

    CN121329484A