Automatic review system and method
By clustering and scoring tender documents through an automated review system and generating a scoring radar chart, the problem of low efficiency in tender document review has been solved, and an efficient and accurate review process has been achieved.
Patent Information
- Application Number
- CN202511910256.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-12-17
- Publication Date
- 2026-05-15
AI Technical Summary
Existing technologies have low efficiency in reviewing tender documents, are prone to errors, and consume a lot of human resources.
An automated review system is adopted, which clusters bid documents through a classification module, establishes a scoring system and generates a scoring radar chart, and uses the scoring radar chart to review the bid documents.
It has enabled automated review of tender documents, improving review efficiency and accuracy, and reducing manual operations and repetitive work.
Smart Images
Figure CN122048260A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of review technology, and more specifically, to an automated review system and method. Background Technology
[0002] Currently, bidding and procurement, as a special procurement method, is widely used in various projects across different industries and fields both domestically and internationally. For example, in the initial bidding process for urban construction projects, a large number of bid documents need to be prepared for participation in bidding project meetings.
[0003] Currently, tender documents are typically drafted manually by staff based on fixed templates. While the general templates are fixed, the specific details are tedious and prone to errors. Therefore, each tender document needs to be reviewed by a designated person from the client, which is time-consuming, inefficient, and prone to overlooking errors. Summary of the Invention
[0004] This invention provides an automated review system and method to solve the problem of low efficiency in bid document review in the prior art, including:
[0005] The classification module is used to obtain the bid text data of the bid documents, cluster each bid document according to the bid text data, and determine the document classification based on the clustering results.
[0006] The system module is used to determine the scoring system for bid documents based on document classification, and to create a scoring radar chart for bid documents based on the scoring system.
[0007] The review module is used to score bid documents based on the scoring radar chart and review the bid documents based on the scoring results.
[0008] Furthermore, the classification module clusters each bid document based on the bid text data, including:
[0009] The bid text data is segmented into words, and the key words in the bid text data are determined based on the segmentation results;
[0010] Calculate the word density of each key word in the bidding text data, and filter out the key words with the highest word density.
[0011] Establish a relevant key word segmentation library, filter out key words with a word segmentation density greater than the first preset threshold, calculate the correlation coefficient between key words with a word segmentation density greater than the first preset threshold and each key word in the relevant key word segmentation library, and take the key words in the relevant key word segmentation library with the largest correlation coefficient as relevant key words;
[0012] The density key words and related key words of the bidding text data are vectorized to obtain density key word vectors and related key word vectors respectively.
[0013] Each bid document is clustered based on its density-based key word segmentation vector and related key word segmentation vectors to obtain the clustering partitions for each bid document.
[0014] Furthermore, the step of clustering each bid document based on density key word segmentation vectors and related key word segmentation vectors to obtain cluster partitions for each bid document includes:
[0015] A sample dataset is established based on the density key word segmentation vectors and related key word segmentation vectors of each tender document, and k initial cluster centers are randomly selected from the sample dataset.
[0016] Calculate the Euclidean distance between the sample data in the sample dataset and the initial cluster center, and cluster the tender documents into the cluster partitions that are closest to the initial cluster centers in terms of Euclidean distance;
[0017] Calculate the mean of the sample data within each cluster partition, and reset the new cluster centers based on the mean of the sample data within each cluster partition;
[0018] Repeat the iteration to get new cluster centers until the cluster centers no longer change or the preset maximum number of iterations is reached, to obtain the cluster partitions of each tender document.
[0019] Furthermore, the system module determines the scoring system for bid documents based on document classification, including:
[0020] Based on historical bidding data, determine the document categories and corresponding scoring systems of historical bidding documents, and establish a training sample set based on the document categories and corresponding scoring systems of historical bidding documents;
[0021] An initial system model is established and trained based on the training sample set to obtain a trained system model.
[0022] The current tender documents are categorized and input into a pre-trained system to build a model, resulting in a corresponding scoring system.
[0023] Furthermore, the system module establishes a scoring radar chart for the bid documents based on the scoring system, including:
[0024] Establish a center point, and draw multiple rays evenly outward from the center point;
[0025] Based on the scoring system, standard key data are determined, and each ray is assigned a standard key data point to obtain the indicator axis, thus generating the scoring radar chart of the tender document.
[0026] Furthermore, the review module scores the bid documents based on the scoring radar chart, including:
[0027] Obtain a second preset threshold and filter out the scoring key words with a word segmentation density greater than the second preset threshold;
[0028] Calculate the cosine similarity between each scoring keyword and each standard key data in the scoring radar chart, and filter out scoring keywords with a cosine similarity greater than the third preset threshold;
[0029] The cosine similarity of the scoring key words with a cosine similarity greater than the third preset threshold is used as an indicator, and each indicator is filled into the indicator axis of the corresponding standard key data.
[0030] Calculate the area value of the scoring radar chart after the indicator axes are filled, and determine the bid score based on the area value of the scoring radar chart after the indicator axes are filled.
[0031] Furthermore, the review module reviews the tender documents based on the scoring results, including:
[0032] Obtain the preset standard bid document score and determine whether the bid document score is greater than the preset standard bid document score;
[0033] If the score of the bid document is greater than the preset standard bid document score, then the corresponding bid document will be determined as the target bid document;
[0034] If the score of a bid document is less than or equal to the score of a pre-set standard bid document, the corresponding bid document will be identified as a non-target bid document.
[0035] On the other hand, the present invention provides an automated review method, executed in a system as described in any of the above embodiments, characterized in that it includes:
[0036] Obtain the bid text data of the bid documents, cluster each bid document based on the bid text data, and determine the document classification based on the clustering results;
[0037] Determine the scoring system for the bid documents based on the document classification, and establish a scoring radar chart for the bid documents based on the scoring system.
[0038] The bid documents are scored based on the scoring radar chart, and the bid documents are reviewed based on the scoring results.
[0039] On the other hand, the present invention provides a computer device, including: a memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the computer instructions to perform the above-described automated review method.
[0040] On the other hand, the present invention provides a computer-readable storage medium, characterized in that the computer-readable storage medium stores computer instructions for causing a computer to perform the above-described automated review method.
[0041] The beneficial effects of this invention are as follows:
[0042] By applying the above technical solutions, this invention establishes a scoring radar chart for each tender document category, and establishes an automated process for tender document review, reducing manual operations and repetitive work, and improving review efficiency and accuracy. Attached Figure Description
[0043] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0044] Figure 1 This invention presents an overall structural diagram of an automated review system proposed in an embodiment of the present invention;
[0045] Figure 2 A flowchart illustrating an automated review method proposed in an embodiment of the present invention is shown. Detailed Implementation
[0046] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.
[0047] This application provides an automated review system, such as... Figure 1 As shown, it includes:
[0048] The classification module is used to acquire the bid text data of the bid documents, cluster the bid documents according to the bid text data, and determine the document classification based on the clustering results; the system module is used to determine the scoring system of the bid documents according to the document classification, and create a scoring radar chart of the bid documents according to the scoring system; the review module is used to score the bid documents according to the scoring radar chart, and review the bid documents according to the scoring results.
[0049] In this embodiment, the bid documents are clustered by bid text data, and the bid documents are clustered by text type. This makes it easier to establish different scoring systems for different types in the future, and to build a scoring radar chart based on the scoring system to conduct a comprehensive and targeted review of each bid document.
[0050] In some embodiments of this application, the classification module clusters each bid document based on the bid text data, including: performing word segmentation on the bid text data, determining key words in the bid text data based on the segmentation results; calculating the word density of each key word in the bid text data, and selecting the density key words with the highest word density; establishing a relevant key word library, selecting key words with a word density greater than a first preset threshold, calculating the correlation coefficient between key words with a word density greater than the first preset threshold and each key word in the relevant key word library, and using the key words in the relevant key word library with the highest correlation coefficient as relevant key words; vectorizing the density key words and relevant key words in the bid text data respectively to obtain density key word vectors and relevant key word vectors; and clustering each bid document based on the density key word vectors and relevant key word vectors to obtain cluster partitions for each bid document.
[0051] In this embodiment, a key word segmentation library is established to store commonly used key words in the bidding documents. Key words with keyword-like characteristics are extracted from the bidding text data through part-of-speech parsing. The word density is calculated based on the frequency of occurrence of each key word in the text data, and density-based key words are selected. A related key word segmentation library is used to store commonly used key words. Then, by calculating the Pearson correlation coefficient, the key words with the highest correlation coefficient are selected as related key words. Density-based key word vectors and related key word vectors are calculated through vectorization, and the bidding documents are clustered based on these vectors.
[0052] In some embodiments of this application, the step of clustering each bid document based on density key word vectors and related key word vectors to obtain cluster partitions for each bid document includes: establishing a sample dataset based on the density key word vectors and related key word vectors of each bid document; randomly selecting k initial cluster centers from the sample dataset; calculating the Euclidean distance between the sample data in the sample dataset and the initial cluster centers, and clustering the bid documents into the cluster partitions with the closest Euclidean distance to the initial cluster centers; calculating the mean of the sample data in each cluster partition, and resetting new cluster centers based on the mean of the sample data in each cluster partition; repeatedly iterating the new cluster centers until the cluster centers no longer change or the preset maximum number of iterations is reached to obtain the cluster partitions for each bid document.
[0053] In this embodiment, the value of k is set to 8, and the tender documents are clustered into 8 cluster partitions.
[0054] In some embodiments of this application, the system module determines the scoring system of the bid documents based on the document classification, including: determining the document classification and corresponding scoring system of historical bid documents based on historical bidding data; establishing a training sample set based on the document classification and corresponding scoring system of historical bid documents; establishing and training an initial system establishment model based on the training sample set to obtain a trained system establishment model; and inputting the document classification of the current bid document into the trained system establishment model to obtain the corresponding scoring system.
[0055] In this embodiment, a training sample set is established by collecting historical bid documents and their corresponding scoring systems from the historical bid review process to train the initial system model, thereby outputting the scoring system corresponding to the cluster partition through the trained system model.
[0056] In some embodiments of this application, the system module establishes a scoring radar chart for the tender document based on the scoring system, including: establishing a center point and drawing multiple rays evenly outward from the center point; determining standard key data based on the scoring system, assigning each ray to a standard key data point to obtain an indicator axis, and thus obtaining the scoring radar chart for the tender document.
[0057] In this embodiment, multiple standard key data indicators are extracted from the scoring system, and each standard key data indicator is mapped onto the corresponding ray of the radar chart, thereby establishing an indicator axis to obtain the scoring radar chart of the tender document.
[0058] In some embodiments of this application, the review module scores the bid documents based on the scoring radar chart, including: obtaining a second preset threshold and filtering out scoring key words with a word density greater than the second preset threshold; calculating the cosine similarity between each scoring key word and each standard key data in the scoring radar chart, and filtering out scoring key words with a cosine similarity greater than a third preset threshold; using the cosine similarity of the scoring key words with a cosine similarity greater than the third preset threshold as an indicator, and filling each indicator into the indicator axis of the corresponding standard key data; calculating the area value of the scoring radar chart after the indicator axis is filled, and determining the bid document score based on the area value of the scoring radar chart after the indicator axis is filled.
[0059] In this embodiment, scoring key words with a word density greater than a second preset threshold are extracted from the bid document. The scoring key words corresponding to each standard key data indicator are selected by calculating the cosine similarity. The cosine similarity between the scoring key words and the corresponding standard key data is filled into the corresponding indicator axis to obtain the area value of the scoring radar chart corresponding to the bid document. The bid document score corresponding to the area value of the scoring radar chart is determined according to the area value-score mapping table. The larger the area value, the higher the corresponding score.
[0060] In some embodiments of this application, the review module reviews the bid documents based on the scoring results, including: obtaining a preset standard bid document score, determining whether the bid document score is greater than the preset standard bid document score; if the bid document score is greater than the preset standard bid document score, then the corresponding bid document is determined as the target bid document; if the bid document score is less than or equal to the preset standard bid document score, then the corresponding bid document is determined as a non-target bid document.
[0061] This embodiment also provides an automated review method, executed in the system described in any of the above embodiments, including:
[0062] S101, Obtain the bid text data of the bid documents, cluster each bid document according to the bid text data, and determine the document classification based on the clustering results;
[0063] S102, Determine the scoring system for the tender documents based on the document classification, and establish a scoring radar chart for the tender documents based on the scoring system;
[0064] S103, score the bid documents according to the scoring radar chart, and review the bid documents based on the scoring results.
[0065] This invention also provides a computer device, including: a memory and a processor, and interfaces for connecting the components, including high-speed interfaces and low-speed interfaces. The components are interconnected via different buses and can be mounted on a common motherboard or otherwise installed as needed. The processor can process instructions executed within the computer device, including instructions stored in or on memory to display graphical information of a GUI on an external input / output device (such as a display device coupled to the interface). In some alternative embodiments, multiple processors and / or multiple buses can be used with multiple memories and multiple memory modules, if desired. Similarly, multiple computer devices can be connected, each providing some of the necessary operations (e.g., as a server array, a group of blade servers, or a multiprocessor system).
[0066] This invention also provides a computer-readable storage medium. The methods described above according to embodiments of the invention can be implemented in hardware or firmware, or implemented as computer code that can be recorded on a storage medium, or implemented as computer code downloaded via a network and originally stored on a remote storage medium or a non-transitory machine-readable storage medium and then stored on a local storage medium. Thus, the methods described herein can be processed by software stored on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. The storage medium can be a magnetic disk, optical disk, read-only memory, random access memory, flash memory, hard disk, or solid-state drive, etc.; further, the storage medium can also include combinations of the above types of memory. It is understood that computers, processors, microprocessor controllers, or programmable hardware include storage components capable of storing or receiving software or computer code, which, when accessed and executed by the computer, processor, or hardware, implements the methods shown in the above embodiments.
[0067] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.
Claims
1. An automated review system, characterized in that, include: The classification module is used to obtain the bid text data of the bid documents, cluster each bid document according to the bid text data, and determine the document classification based on the clustering results. The system module is used to determine the scoring system for bid documents based on document classification, and to create a scoring radar chart for bid documents based on the scoring system. The review module is used to score bid documents based on the scoring radar chart and review the bid documents based on the scoring results.
2. The automated review system according to claim 1, characterized in that, The classification module clusters each bid document based on the bid text data, including: The bid text data is segmented into words, and the key words in the bid text data are determined based on the segmentation results; Calculate the word density of each key word in the bidding text data, and filter out the key words with the highest word density. Establish a relevant key word segmentation library, filter out key words with a word segmentation density greater than the first preset threshold, calculate the correlation coefficient between key words with a word segmentation density greater than the first preset threshold and each key word in the relevant key word segmentation library, and take the key words in the relevant key word segmentation library with the largest correlation coefficient as relevant key words; The density key words and related key words of the bidding text data are vectorized to obtain density key word vectors and related key word vectors respectively. Each bid document is clustered based on its density-based key word segmentation vector and related key word segmentation vectors to obtain the clustering partitions for each bid document.
3. The automated review system according to claim 2, characterized in that, The process of clustering each bid document based on density-based key word segmentation vectors and related key word segmentation vectors to obtain cluster partitions for each bid document includes: A sample dataset is established based on the density key word segmentation vectors and related key word segmentation vectors of each tender document, and k initial cluster centers are randomly selected from the sample dataset. Calculate the Euclidean distance between the sample data in the sample dataset and the initial cluster center, and cluster the tender documents into the cluster partitions that are closest to the initial cluster centers in terms of Euclidean distance; Calculate the mean of the sample data within each cluster partition, and reset the new cluster centers based on the mean of the sample data within each cluster partition; Repeat the iteration to get new cluster centers until the cluster centers no longer change or the preset maximum number of iterations is reached, to obtain the cluster partitions of each tender document.
4. The automated review system according to claim 1, characterized in that, The system module determines the scoring system for bid documents based on document classification, including: Based on historical bidding data, determine the document categories and corresponding scoring systems of historical bidding documents, and establish a training sample set based on the document categories and corresponding scoring systems of historical bidding documents; An initial system model is established and trained based on the training sample set to obtain a trained system model. The current tender documents are categorized and input into a pre-trained system to build a model, resulting in a corresponding scoring system.
5. The automated review system according to claim 2, characterized in that, The system module establishes a scoring radar chart for the bid documents based on the scoring system, including: Establish a center point, and draw multiple rays evenly outward from the center point; Based on the scoring system, standard key data are determined, and each ray is assigned a standard key data point to obtain the indicator axis, thus generating the scoring radar chart of the tender document.
6. The automated review system according to claim 5, characterized in that, The review module scores the bid documents based on the scoring radar chart, including: Obtain a second preset threshold and filter out the scoring key words with a word segmentation density greater than the second preset threshold; Calculate the cosine similarity between each scoring keyword and each standard key data in the scoring radar chart, and filter out scoring keywords with a cosine similarity greater than the third preset threshold; The cosine similarity of the scoring key words with a cosine similarity greater than the third preset threshold is used as an indicator, and each indicator is filled into the indicator axis of the corresponding standard key data. Calculate the area value of the scoring radar chart after the indicator axes are filled, and determine the bid score based on the area value of the scoring radar chart after the indicator axes are filled.
7. The automated review system according to claim 6, characterized in that, The review module reviews the bid documents based on the scoring results, including: Obtain the preset standard bid document score and determine whether the bid document score is greater than the preset standard bid document score; If the score of the bid document is greater than the preset standard bid document score, then the corresponding bid document will be determined as the target bid document; If the score of a bid document is less than or equal to the score of a pre-set standard bid document, the corresponding bid document will be identified as a non-target bid document.
8. An automated review method, executed in the system as described in any one of claims 1-7, characterized in that, include: Obtain the bid text data of the bid documents, cluster each bid document based on the bid text data, and determine the document classification based on the clustering results; Determine the scoring system for the bid documents based on the document classification, and establish a scoring radar chart for the bid documents based on the scoring system. The bid documents are scored based on the scoring radar chart, and the bid documents are reviewed based on the scoring results.
9. A computer device, characterized in that, include: The system includes a memory and a processor, which are interconnected. The memory stores computer instructions, and the processor executes the computer instructions to perform the automated review method of claim 8.
10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions for causing the computer to execute the automated review method of claim 8.