Bidding risk early warning method and system based on artificial intelligence

Abstract

The invention discloses a bidding and tendering risk early warning method and system based on artificial intelligence, and relates to the field of artificial intelligence risk assessment. Bidding and tendering related data and bidding and tendering enterprise credit data are collected from multiple angles and multiple sources, and are stored in a bidding and tendering database after format alignment is completed; calling bidding and tendering data in the bidding and tendering database, generating a text data set, analyzing financial data based on the text data set, extracting trend characteristics, and calculating a financial ratio; cyclically traversing the text data set, and obtaining keywords and key sentences with bidding and tendering relevance higher than a preset threshold in the text data set through text similarity and keyword feature extraction; and constructing a risk prediction model, inputting the financial ratio, the keywords and the key sentences into the risk prediction model, outputting an early warning value, and determining a current bidding and tendering risk level. According to the invention, the accurate threshold value is preset according to the project and the history, and the evaluation data is updated in real time, thereby realizing real-time accurate early warning to assist related parties in reducing loss.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence risk assessment, and more specifically, to a bidding risk early warning method and system based on artificial intelligence. Background Art

[0002] For a long time, the bidding risk early warning work mainly relied on relevant personnel to review bidding documents and processes based on their own experience. This method not only required a large amount of manpower and time, but also had strong subjectivity. Different reviewers had different judgment criteria and sensitivities to risks, which easily led to the lack of consistency and accuracy in early warning results. In addition, traditional methods had limited capabilities in data analysis, making it difficult to comprehensively and deeply analyze a large amount of bidding data, unable to discover the underlying correlation relationships and potential risks hidden behind the data, and thus difficult to build an effective risk assessment model, resulting in the inability to timely and accurately detect risks. Moreover, the traditional risk early warning mechanism could not monitor bidding activities in real time, but only conducted reviews and processing afterwards. Facing sudden and time-sensitive risks, it was difficult to take effective response measures in the first time, thus causing serious losses.

[0003] In recent years, artificial intelligence technology has made breakthrough progress. Technologies such as machine learning, deep learning, and natural language processing can quickly and accurately analyze a large amount of bidding data, and discover the underlying laws and risk characteristics therein. By constructing an intelligent risk early warning model, real-time monitoring and dynamic early warning of bidding activities can be realized, and various risks can be discovered and prevented in a timely manner. Applying artificial intelligence technology to the field of bidding risk early warning can not only significantly improve the accuracy and efficiency of risk early warning, reduce the interference of human factors, but also provide scientific decision-making basis for regulatory authorities and enterprises, and help the healthy and orderly development of the bidding market.

[0004] Therefore, researching and developing a bidding risk early warning method and system based on artificial intelligence has extremely important practical significance and application value, and is expected to break through the limitations of the traditional risk early warning mode and greatly improve the risk management level of bidding activities. Summary of the Invention

[0005] In view of this, the present invention provides a bidding risk early warning method and system based on artificial intelligence to solve the problems existing in the background art.

[0006] To achieve the above object, the present invention adopts the following technical solutions:

[0007] A bidding risk early warning method based on artificial intelligence, comprising the following steps:

[0008] Connect to the bidding platform, enterprise official website and credit database, collect bidding-related data and credit data of bidding enterprises from multiple perspectives and multiple sources, store them in the bidding database after format alignment;

[0009] Retrieve the bidding data in the bidding database, which is divided into non-text data and text data. Convert the non-text data into text data to generate a text dataset. Analyze the financial data based on the text dataset to extract trend features and calculate financial ratios;

[0010] Loop through the text dataset, and obtain keywords and key sentences in the text dataset with a bidding relevance higher than the preset threshold through text similarity and keyword feature extraction;

[0011] Preset the warning threshold based on the project nature scale and historical risk data, collect and update data in real time, construct a risk prediction model, input the financial ratios, keywords and key sentences into the risk prediction model, output the warning value, and compare the warning value with the preset warning threshold to determine the current bidding risk level.

[0012] Optionally, convert the non-text data into text data. Specifically: traverse and retrieve the non-text data in turn, excavate the structured data in the non-text data, and perform preprocessing and standardization processing on the structured data; based on the standardized structured data, generate picture text, and use optical character recognition technology to obtain the text data in the picture text to complete the data conversion.

[0013] Optionally, analyze the financial data based on the text dataset to extract trend features and calculate financial ratios. Specifically: use natural language processing tools to extract financial key performance indicators, extract trend features by analyzing the time series features of financial key performance indicators; calculate financial ratios according to the development ability ratios.

[0014] Optionally, loop through the text dataset, and obtain keywords and key sentences in the text dataset with a bidding relevance higher than the preset threshold through text similarity and keyword feature extraction. Specifically, it includes the following steps:

[0015] Loop through each piece of bidding data in the text dataset, compare the similarity of each piece of bidding data with the remaining bidding data. When the similarity is greater than the preset threshold, give a warning of the possibility of bid rigging;

[0016] If the similarity is less than the preset threshold, use the keyword feature extraction method to extract the keywords of each piece of bidding data, and locate the key sentences where the keywords are located to obtain the pre-screened keywords and key sentences;

[0017] Calculate the first correlation value between the keyword and the current bidding information. When the first correlation value is greater than the first threshold, determine that the keyword and the key sentence where the keyword is located are bidding information. When the correlation value is less than the first threshold, calculate the second correlation value between the key sentence where the keyword is located and the current bidding information. When the second correlation value is greater than the second threshold, determine that the current key sentence is bidding information.

[0018] Optionally, keyword feature extraction includes the following steps:

[0019] Use clustering algorithm to extract initial keywords from data text, analyze current initial keywords according to bidding statement preference, and determine word weight by calculating word frequency and inverse document frequency in text;

[0020] The initial keywords are represented based on the principal component analysis algorithm, and the keyword features are extracted by weighted averaging, feature-level fusion, and decision-level fusion.

[0021] Optionally, when a warning of the possibility of bid collusion occurs, it is sent to the relevant parties and the current bidding data is eliminated.

[0022] Optionally, a risk prediction model is constructed using a BP neural network, and the risk prediction model is optimized using a loss function.

[0023] An artificial intelligence-based bidding risk early warning system, comprising:

[0024] Bidding data collection and storage module: used to connect the bidding platform, the company's official website and the credit database, collect bidding-related data and bidding company credit data from multiple angles and sources, and store them in the bidding database after format alignment;

[0025] Bidding data analysis module: used to retrieve bidding data from the bidding database, divide it into non-text data and text data, convert non-text data into text data, generate text data sets, analyze financial data based on the text data sets to extract trend features, and calculate financial ratios;

[0026] Keyword and key sentence acquisition module: used to loop through the text data set and obtain the keywords and key sentences in the text data set whose bidding relevance is higher than the preset threshold by extracting text similarity and keyword features;

[0027] Bidding risk level determination module: used to preset warning thresholds based on the nature and scale of the project and historical risk data, collect and update data in real time, build a risk prediction model, input financial ratios, keywords and key sentences into the risk prediction model, output warning values, compare the warning values with the preset warning thresholds, and determine the current bidding risk level.

[0028] It can be seen from the above technical solutions that, compared with the prior art, the present invention provides a bidding risk warning method and system based on artificial intelligence, which has the following beneficial effects:

[0029] 1. By connecting the bidding platform, the company's official website and the credit database, data is collected from multiple angles and sources, which can comprehensively cover various information related to bidding, including information on the project itself and the company's credit data, avoiding the problem of incomplete information that may exist in a single data source, making risk assessment more accurate and comprehensive.

[0030] 2. Analyzing financial data based on text data sets to extract trend features and calculate financial ratios can provide an in-depth understanding of the company's operating conditions and potential risks from a financial perspective, providing quantitative indicators for risk assessment. At the same time, by extracting text similarity and keyword features to obtain keywords and key sentences with high relevance to bidding, it is possible to capture key information related to bidding risks in text data and mine potential risks from a semantic level.

[0031] 3. Real-time data collection and update and construction of risk prediction models can timely reflect the dynamic changes of the market and projects and ensure the timeliness of risk assessment. Financial ratios, keywords and key sentences are input into the risk prediction model to output warning values, and compared with the preset warning thresholds to determine the risk level, thus achieving real-time monitoring and accurate assessment of bidding risks, which helps relevant parties to take timely measures to deal with risks and reduce losses. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.

[0033] Figure 1 It is a schematic diagram of the method flow of the present invention;

[0034] Figure 2 It is a schematic diagram of the system structure of the present invention. DETAILED DESCRIPTION

[0035] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0036] An embodiment of the present invention discloses a bidding risk early warning method based on artificial intelligence, as Figure 1 shown, which includes the following steps:

[0037] Step 1: Connect to the bidding platform, enterprise official website and credit database, collect bidding-related data and bidding enterprise credit data from multiple angles and multiple sources, and store them in the bidding database after format alignment;

[0038] Step 2: Retrieve the bidding data in the bidding database, which is divided into non-text data and text data. Convert the non-text data into text data to generate a text data set. Analyze the financial data based on the text data set to extract trend features and calculate financial ratios;

[0039] Step 3: Traverse the text data set in a loop, and obtain keywords and key sentences in the text data set with a bidding relevance higher than a preset threshold through text similarity and keyword feature extraction;

[0040] Step 4: Preset an early warning threshold according to the project nature scale and historical risk data, collect and update data in real time, construct a risk prediction model, input the financial ratios, keywords and key sentences into the risk prediction model, output an early warning value, and compare the early warning value with the preset early warning threshold to determine the current bidding risk level.

[0041] Further, in Step 2, converting the non-text data into text data specifically includes: sequentially traversing and retrieving the non-text data, mining the structured data in the non-text data, and performing preprocessing and standardization processing on the structured data; generating picture text based on the standardized structured data, and using optical character recognition technology to obtain the text data in the picture text to complete the data conversion.

[0042] Furthermore, analyzing the financial data based on the text data set in Step 2 to extract trend features and calculate financial ratios specifically includes: using natural language processing tools to extract financial key performance indicators, extracting trend features by analyzing the time series features of the financial key performance indicators; calculating financial ratios according to the development ability ratios.

[0043] Further, in Step 3, traversing the text data set in a loop and obtaining keywords and key sentences in the text data set with a bidding relevance higher than a preset threshold through text similarity and keyword feature extraction specifically includes the following steps:

[0044] Step 3.1: Traverse each piece of bidding data in the text dataset in a loop, compare each piece of bidding data with the remaining bidding data for similarity. When the similarity is greater than the preset threshold, a warning of possible bid rigging is given. First, convert the bidding text data into vector form through a word vector model, and then calculate the cosine value of the cosine angle between the two vectors. The closer this value is to 1, the higher the text similarity. Measure the minimum number of single-character edit operations (insertion, deletion, replacement) required to convert one string into another string. The smaller the edit distance, the higher the text similarity.

[0045] Step 3.2: If the similarity is less than the preset threshold, use the keyword feature extraction method to extract the keywords of each piece of bidding data and locate the key sentences where the keywords are located, obtaining the pre-screened keywords and key sentences.

[0046] Step 3.3: Calculate the first relevance value of the keyword to the current bidding information. When the first relevance value is greater than the first threshold, determine the keyword and the key sentence where the keyword is located as the bidding information. When the relevance value is less than the first threshold, calculate the second relevance value of the key sentence where the keyword is located to the current bidding information again. When the second relevance value is greater than the second threshold, determine the current key sentence as the bidding information.

[0047] Furthermore, in Step 3.1, a nested loop structure is used to process each piece of bidding data in the dataset. The outer loop is used to select a reference piece of bidding data, and the inner loop is responsible for comparing the reference data with each of the remaining pieces of bidding data.

[0048] Furthermore, the keyword feature extraction includes the following steps:

[0049] Step 3.2.1: Use a clustering algorithm to extract the initial keywords in the data text, analyze the current initial keywords according to the preference of bidding statements, and determine the weight of the words by calculating the word frequency and inverse document frequency of the words in the text.

[0050] Step 3.2.2: Based on the principal component analysis algorithm, perform feature representation on the initial keywords, and select weighted average, feature-level fusion, and decision-level fusion to complete keyword feature extraction.

[0051] Further, when a warning of possible bid rigging appears, it is sent to the relevant parties, and the current bidding data is excluded.

[0052] Further, in Step 4, use a BP neural network to construct a risk prediction model, and optimize the risk prediction model using a loss function. Among them, the preset warning threshold specifically includes the following two steps:

[0053] Based on historical data statistics: Statistically analyze historical risk data, and calculate statistics such as the average value and standard deviation of different types of projects under different risk indicators. According to these statistics, combined with the acceptable risk level of the project, set reasonable warning thresholds. For example, for the cost overrun risk of a certain type of project, if the historical data shows that the average overrun rate is 10% and the standard deviation is 3%, the mild warning threshold can be set when the cost overrun rate reaches 13% (average value + standard deviation), and the severe warning threshold can be set when it reaches 16% (average value + 2 times the standard deviation).

[0054] Consider the particularity of the project: When setting the threshold, fully consider the particularity of the current project, such as factors like project urgency, technical difficulty, and market competition situation. For projects with higher technical difficulty, it may be necessary to appropriately lower the risk warning threshold to detect and respond to potential risks earlier.

[0055] Furthermore, for the optimization of the risk prediction model, as real-time data is continuously collected, new data can be added to the training dataset in a timely manner, and the model can be retrained regularly to adapt to changes in the market environment and project characteristics. Regularly evaluate the performance of the model to check whether the prediction effect of the model on new data is still good. If it is found that the model performance deteriorates, analyze the reasons in a timely manner and adjust the model, such as updating model parameters, adding new features, or reselecting the model, etc.

[0056] And Figure 1 Corresponding to the method shown above, the present invention also discloses an artificial intelligence-based bidding risk warning system for Figure 1 the implementation of the method, and the specific structure is as Figure 2 shown, including:

[0057] Bidding data collection and storage module: Used to connect to the bidding platform, enterprise official website and credit database, collect bidding-related data and bidding enterprise credit data from multiple angles and multiple sources, and store them in the bidding database after format alignment;

[0058] Bidding data analysis module: Used to retrieve the bidding data in the bidding database, which is divided into non-text data and text data, convert the non-text data into text data, generate a text dataset, analyze the financial data based on the text dataset to extract trend features, and calculate financial ratios;

[0059] Keyword and key sentence acquisition module: Used to loop through the text dataset, and obtain keywords and key sentences in the text dataset with a bidding relevance higher than the preset threshold through text similarity and keyword feature extraction;

[0060] Bidding and tendering risk level determination module: used to preset early warning thresholds based on the nature and scale of the project and historical risk data, collect and update data in real time, construct a risk prediction model, input financial ratios, keywords and key sentences into the risk prediction model, output early warning values, compare the early warning values with the preset early warning thresholds, and determine the current bidding and tendering risk level.

[0061] The various embodiments in this specification are described in a progressive manner. Each embodiment focuses on the differences from other embodiments. For the same or similar parts among the various embodiments, reference can be made to each other. For the devices disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the description is relatively simple. For the relevant parts, reference can be made to the description in the method section.

[0062] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present invention. Various modifications to these embodiments will be obvious to those skilled in the art. The general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but will be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. An artificial intelligence-based bidding risk early warning method, characterized in that, It includes the following steps: Connect to the bidding platform, enterprise official website and credit database, collect bidding-related data and bidding enterprise credit data from multiple perspectives and multiple sources, and store them in the bidding database after format alignment; Retrieve the bidding data in the bidding database, which is divided into non-text data and text data. Convert the non-text data into text data to generate a text data set. Analyze the financial data based on the text data set to extract trend features and calculate financial ratios; Loop through the text data set, and obtain the keywords and key sentences in the text data set with a bidding relevance higher than the preset threshold through text similarity and keyword feature extraction; Preset the warning threshold according to the project nature scale and historical risk data, collect and update the data in real time, construct a risk prediction model, input the financial ratios, keywords and key sentences into the risk prediction model, output the warning value, compare the warning value with the preset warning threshold, and determine the current bidding risk level.

2. The method for warning of bidding risks based on artificial intelligence according to claim 1, wherein, Convert the non-text data into text data, specifically: traverse and retrieve the non-text data in turn, excavate the structured data in the non-text data, and perform preprocessing and standardization processing on the structured data; based on the standardized structured data, generate picture text, and use optical character recognition technology to obtain the text data in the picture text to complete the data conversion.

3. The method for warning of bidding risks based on artificial intelligence according to claim 1, wherein Analyze the financial data based on the text data set to extract trend features and calculate financial ratios, specifically: use natural language processing tools to extract financial key performance indicators, and extract trend features by analyzing the time series features of the financial key performance indicators; Calculate the financial ratio according to the development ability ratio.

4. The method for warning of bidding risks based on artificial intelligence according to claim 1, characterized in that Loop through the text data set, and obtain the keywords and key sentences in the text data set with a bidding relevance higher than the preset threshold through text similarity and keyword feature extraction, specifically including the following steps: Loop through each bidding data in the text data set, compare the similarity of each bidding data with the remaining bidding data, and when the similarity is greater than the preset threshold, give a warning of the possibility of bid rigging; If the similarity is less than the preset threshold, use the keyword feature extraction method to extract the keywords of each bidding data and locate the key sentences where the keywords are located to obtain the pre-screened keywords and key sentences; Calculate the first relevance value between the keyword and the current bidding information. When the first relevance value is greater than the first threshold, determine that the keyword and the key sentence where the keyword is located are bidding information. When the relevance value is less than the first threshold, calculate the second relevance value between the key sentence where the keyword is located and the current bidding information again. When the second relevance value is greater than the second threshold, determine that the current key sentence is bidding information.

5. The method for warning of bidding risks based on artificial intelligence according to claim 3, characterized in that, The keyword feature extraction includes the following steps: Use the clustering algorithm to extract the initial keywords in the data text, analyze the current initial keywords according to the bidding statement preference, and determine the weight of the words by calculating the word frequency and inverse document frequency of the words in the text; Perform feature representation on the initial keywords based on the principal component analysis algorithm, and select weighted average, feature-level fusion, and decision-level fusion to complete keyword feature extraction.

6. The method for warning of bidding risks based on artificial intelligence according to claim 3, wherein When a warning of the possibility of bid rigging appears, it is sent to the relevant parties to exclude the current bidding and tendering data.

7. The method for bidding risk early warning based on artificial intelligence according to claim 1, characterized in that Use the BP neural network to construct a risk prediction model, and use the loss function to optimize the risk prediction model.

8. An artificial intelligence-based bidding risk early warning system, characterized in that, Including: Bidding and tendering data collection and storage module: used to connect with the bidding and tendering platform, enterprise official website and credit database, collect bidding and tendering related data and bidding and tendering enterprise credit data from multiple angles and multiple sources, and store them in the bidding and tendering database after format alignment; Bidding and tendering data analysis module: used to retrieve the bidding and tendering data in the bidding and tendering database, which is divided into non-text data and text data, convert the non-text data into text data, generate a text data set, analyze the financial data based on the text data set to extract trend features, and calculate financial ratios; Keyword and key sentence acquisition module: used to loop through the text data set, and obtain keywords and key sentences in the text data set with a bidding and tendering relevance higher than the preset threshold through text similarity and keyword feature extraction; Bidding and tendering risk level determination module: used to preset the warning threshold according to the project nature scale and historical risk data, collect and update data in real time, construct a risk prediction model, input the financial ratios, keywords and key sentences into the risk prediction model, output the warning value, compare the warning value with the preset warning threshold, and determine the current bidding and tendering risk level.

Images