Intelligent quality control method and system for engineering cost file based on multi-dimensional data

Abstract

The present application belongs to the technical field of data processing, and particularly relates to an engineering cost file intelligent quality control method and system based on multidimensional data. The method comprises the following steps: analyzing a cost file to obtain project characteristic text, comprehensive unit price and engineering quantity of a target cost item, and searching for a standard reference item containing standard reference text, standard unit price distribution data and scale reference engineering quantity; determining text variability based on the number of missing core process characteristic words and string edit distance; calculating the degree of deviation of comprehensive unit price from standard unit price distribution data and adjusting using engineering quantity ratio to obtain price and quantity disturbance energy; adjusting the price and quantity disturbance energy using the difference between the text tolerance boundary parameter and the text variability to obtain an abnormality index, and outputting an intelligent quality control diagnosis report when the index exceeds a threshold. The present application can break through the limitation of single-dimensional comparison, accurately identify multi-dimensional coupled hidden bid defects, and effectively reduce false positives and false negatives.

Description

Technical Field

[0001] This invention relates to the field of data processing technology. More specifically, this invention relates to an intelligent quality control method and system for engineering cost documents based on multi-dimensional data. Background Technology

[0002] With the deepening of the digital transformation of engineering cost, the automatic quality control of engineering cost documents has become a core link in ensuring the investment benefits of projects. In actual business processing scenarios, engineering cost documents often contain a large amount of pricing item data, and each item covers multi-dimensional core information such as project characteristic text, quantity of work and comprehensive unit price.

[0003] Currently, in the field of cost quality control, the traditional method based on static rule bases and benchmark comparison is widely used. This method mainly extracts the comprehensive unit price from the cost document and directly compares it with the standard unit price range of similar projects in the historical database. If the value exceeds the set range, an alarm is triggered.

[0004] However, in actual engineering project operations, cost data exhibits strong multi-dimensional coupling characteristics. Construction companies often employ covert unbalanced pricing strategies, creating the appearance of non-standard projects by subtly replacing or modifying project descriptions to conceal high unit prices; or by hiding minor price increases in core items with large amounts of basic work. Existing technologies use isolated and static review logic, completely severing the inherent correlation between textual variations, project volume, and unit price fluctuations. This single-dimensional comparison mechanism is highly susceptible to failure due to the rigid constraints of static boundaries when dealing with complex scenarios where slight textual disturbances accompany price deviations. This leads to a significant decrease in detection sensitivity, allowing numerous abnormal pricing items with deep logical distortions to easily evade review, resulting in frequent underreporting and impacting the overall financial security and cost control effectiveness of the project. Summary of the Invention

[0005] To address the technical problem that the existing single comparison mechanism cannot identify hidden price anomalies with multidimensional coupling and is prone to false negatives, the present invention provides solutions in the following aspects.

[0006] In a first aspect, the present invention provides an intelligent quality control method for engineering cost documents based on multi-dimensional data, comprising: Parse the engineering cost document to obtain the target cost item, and extract the project feature text, comprehensive unit price, quantity and bill of quantities classification code of the target cost item; according to the bill of quantities classification code, retrieve the matching standard reference item from the historical benchmark database, and extract the standard reference text, standard unit price distribution data and scale reference quantity corresponding to the standard reference item. Extract core process feature words from the project feature text and standard reference text. Based on the number of missing core process feature words and the string edit distance between the project feature text and the standard reference text, determine the text variability of the target cost item. The degree of deviation of the comprehensive unit price from the standard unit price distribution data is calculated, and the degree of deviation is adjusted by the ratio of the quantity of work to the scale reference quantity of work, so as to obtain the price-quantity disturbance energy of the target cost item; The price perturbation energy is adjusted based on the difference between the preset text tolerance boundary parameter and the text variability to obtain the anomaly index of the target cost item. In response to an abnormal index exceeding the warning threshold, the target cost item is marked as a high-risk item and an intelligent quality control diagnostic report is output.

[0007] Preferably, retrieving matching standard reference entries from the historical benchmark database based on the list classification code includes: Based on the list classification code, retrieve from the historical benchmark database the item that belongs to the same category as the target cost item and has the smallest difference in overall project volume, and use it as the standard reference item.

[0008] Preferably, the extraction of core process feature words from the project feature text and standard reference text includes: By using word segmentation algorithms and engineering-specific dictionaries, the project feature text and standard reference text are segmented and parsed to extract several core process feature words contained in the project feature text and standard reference text.

[0009] Preferably, determining the textual variability of the target cost item includes: The number of core process feature words in the standard reference text that are present but not covered by the core process feature words of the target cost item is counted to obtain the number of missing feature words; the edit distance between the project feature text and the standard reference text is calculated, and the maximum character length of the project feature text and the standard reference text is obtained. The edit distance is normalized using the maximum character length to obtain the distance deviation; a feature missing penalty term is constructed based on the ratio between the number of missing feature words and the total number of core process feature words in the standard reference text; and the text variability is determined by combining the feature missing penalty term and the distance deviation.

[0010] Preferably, the step of calculating the deviation of the comprehensive unit price from the standard unit price distribution data, and adjusting the deviation using the ratio of the quantity of work to the scale reference quantity of work, to obtain the price-quantity disturbance energy of the target cost item, includes: Extract the standard unit price mean and standard unit price standard deviation from the standard unit price distribution data, calculate the difference between the comprehensive unit price and the standard unit price mean, extract the absolute value of the calculation result and divide it by the standard unit price standard deviation to obtain the basic price deviation; perform natural exponential mapping on the negative number of the ratio of the quantity of work to the scale reference quantity of work to obtain the attenuation median value, and use the difference between the value 1 and the attenuation median value as the quantity of work release coefficient; multiply the basic price deviation by the quantity of work release coefficient to obtain the price and quantity disturbance energy.

[0011] Preferably, the adjustment of the price disturbance energy based on the difference between a preset text tolerance boundary parameter and text variability to obtain the anomaly index of the target cost item includes: The maximum value between the text variability and the preset minimum positive constant is obtained as the anti-overflow denominator; the basic disturbance term is determined based on the ratio of the price-quantity disturbance energy to the anti-overflow denominator; a nonlinear dynamic damping term is constructed using the hyperbolic tangent function based on the difference between the text tolerance boundary parameter and the text variability; the nonlinear dynamic damping term is used to adjust the basic disturbance term through product to obtain the anomaly index of the target cost item.

[0012] Preferably, the output intelligent quality control diagnostic report includes: Extract the project feature text, comprehensive unit price, quantity, list classification code, text variability, and price-quantity disturbance energy of all high-risk items, summarize and splice them to generate an intelligent quality control diagnostic report for engineering cost documents and output it.

[0013] Secondly, the present invention provides an intelligent quality control system for engineering cost documents based on multi-dimensional data, including a processor and a memory. The memory stores computer program instructions, and when the computer program instructions are executed by the processor, the above-mentioned intelligent quality control method for engineering cost documents based on multi-dimensional data is implemented.

[0014] By adopting the above technical solution, the intelligent quality control method for engineering cost documents based on multi-dimensional data is generated into a computer program and stored in a memory so that it can be loaded and executed by a processor. This allows for the creation of a terminal device based on the memory and processor, making it convenient to use.

[0015] The beneficial effects of this invention are as follows: This invention abandons the single-dimensional static judgment mode that relies solely on fixed boundary comparisons of comprehensive unit prices. It determines text variability by analyzing the number of missing core process feature words and string edit distance, enabling in-depth analysis of changes in text topology and objectively measuring the actual degree of change in project characteristics. This effectively identifies non-standard project appearances created through micro-level word substitution. This invention calculates the degree of deviation of the comprehensive unit price from the standard unit price distribution data and adjusts the deviation degree using the ratio of the project quantity to the scale reference project quantity to obtain the price-quantity disturbance energy. By combining the discrete state of the price with the scale distribution of the project quantity for joint calculation, it amplifies the impact weight of small price deviations on the total cost under massive project quantities, while suppressing abnormal price fluctuations under extremely small project quantities, truly reflecting the real fluctuation energy hidden in the cost data. This invention utilizes a nonlinear dynamic damping term constructed based on preset text tolerance boundary parameters and the difference in text variability to adjust the energy of price and quantity disturbances, obtaining an anomaly index. It constructs a reverse collaborative verification mechanism between text deviation and price disturbances: when the unit price of a cost item is adjusted without significant textual changes, the system amplifies its anomalous characteristics due to a lack of reasonable justification for the change; conversely, when the unit price change is supported by sufficient semantic alteration, the system attenuates and tolerates its anomaly index. This invention's shift from single-value review to multi-dimensional logical deduction aligns with the inherent physical logic of cost auditing, overcomes the evasion of static review boundaries by anomalous items, effectively eliminates false alarms caused by reasonable changes while identifying and intercepting hidden unbalanced pricing behavior, and improves the detection sensitivity and judgment reliability of intelligent quality control of engineering cost documents. Attached Figure Description

[0016] Figure 1 This is a flowchart illustrating the intelligent quality control method for engineering cost documents based on multi-dimensional data in this invention; Figure 2 This is a schematic diagram of the abnormal index distribution in an embodiment of the present invention; Figure 3 This is a schematic diagram illustrating the mapping relationship between text variability and anomaly index in an embodiment of the present invention; Figure 4 This is a schematic diagram comparing the unit price deviation of the traditional method with the anomaly index of the present invention. Detailed Implementation

[0017] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] The specific embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0019] This invention discloses an intelligent quality control method for engineering cost documents based on multi-dimensional data, referring to... Figure 1 This includes steps S1-S5: S1. Parse the engineering cost document to obtain the target cost item, and extract the project feature text, comprehensive unit price, quantity and list classification code of the target cost item; according to the list classification code, retrieve the matching standard reference item from the historical benchmark database, extract the standard reference text, standard unit price distribution data and scale reference quantity corresponding to the standard reference item, and extract the project feature text and the core process feature words in the standard reference text.

[0020] Specifically, natural language processing and table structure parsing techniques are used to parse the input engineering cost document to obtain several target cost items containing underlying pricing information.

[0021] For any target cost item, extract its project feature text, comprehensive unit price, quantity, and list classification code used to identify project attributes.

[0022] The project feature text was segmented and parsed using a Chinese automatic word segmentation algorithm (jieba) and a professional dictionary in the engineering field, and several core process feature words were extracted from the project feature text.

[0023] Simultaneously, based on the list classification code, the system retrieves the standard reference item from the historical benchmark database that belongs to the same category as the target cost item and has the closest overall project volume. It then extracts the corresponding standard reference text, standard unit price distribution data (including the standard unit price mean and standard unit price standard deviation), and scale reference quantities. Similarly, it uses an automatic Chinese word segmentation algorithm and a professional engineering dictionary to segment and parse the standard reference text, extracting several core process feature words contained within it.

[0024] S2. Based on the number of missing core process feature words and the string edit distance between the project feature text and the standard reference text, determine the text variability of the target cost item.

[0025] It should be noted that minor changes to project feature text are often used as a reason to change the pricing benchmark, making it impossible to identify unreasonable quotations concealed by textual modifications through a single numerical comparison. Therefore, this invention obtains textual variability by comparing the character topological deviation and core process omissions between the target cost item and the standard reference item at the text level, providing a reference for subsequently judging the rationality of price fluctuations.

[0026] Specifically, the number of core process feature words in the standard reference text that are not covered by the core process feature words of the target cost item is counted to obtain the number of missing feature words. The Levenshtein distance algorithm is used to calculate the edit distance between the project feature text and the corresponding standard reference text.

[0027] Extract the character length of the project feature text and the character length of the standard reference text. Determine the text variability of the target cost item based on the maximum value of the two, the number of missing feature words, and the edit distance. In the formula, Indicates the first Textual variability of each target cost item; Indicates the first Edit distance between the project feature text and the corresponding standard reference text of each target cost item; Indicates the first The maximum character length between the project feature text and the corresponding standard reference text for each target cost item is used to determine the edit distance. Perform absolute value normalization; This indicates that the standard reference text exists but is not specified in the first reference. The number of missing feature words covered by the core process feature words of each target cost item; Indicates the first The total number of core process feature words contained in the standard reference text corresponding to each target cost item; This represents an exponential function with the natural constant as its base. To prevent computational overflow, when... , When the value is 0, the system assigns it a preset minimum positive constant. This ensures that the calculation remains legally closed-loop even under any extreme engineering cost data. In this embodiment, In other embodiments, the setting is 1, and implementers can adjust it according to the actual implementation situation. .

[0028] When the number of missing feature words Larger, edit distance The larger the value, the more profound the modification of the project feature text to the corresponding standard reference text, and the greater the textual variability of the target cost item. The larger the value, the more accurately the present invention can assess the degree of text change through text variability.

[0029] S3. Calculate the degree of deviation of the comprehensive unit price from the standard unit price distribution data, and adjust the degree of deviation using the ratio of the quantity of work to the scale reference quantity of work to obtain the price and quantity disturbance energy of the target cost item.

[0030] It should be noted that since the absolute deviation of the unit price cannot fully reflect the risk of the total cost, the difference in the impact of the same price deviation on the total cost under extremely small and massive project quantities is completely ignored. Therefore, this invention combines the dispersion of the comprehensive unit price and the standard unit price with the scale distribution characteristics of the project quantity to perform joint calculation to obtain the price and quantity disturbance energy, thereby truly and accurately reflecting the magnitude of the abnormal fluctuation energy hidden in the cost data.

[0031] Specifically, the price-quantity disturbance energy of the target cost item satisfies the expression: In the formula, Indicates the first The price-quantity disturbance energy of each target cost item; Indicates the first The comprehensive unit price of each target cost item; Indicates the first The average standard unit price of the standard reference item corresponding to each target cost item; Indicates the first Standard unit price standard deviation of the standard reference item corresponding to each target cost item; This indicates the absolute value operation; Indicates the first The quantity of work for each target cost item; Indicates the first The scale and reference quantity of the standard reference item corresponding to each target cost item; This represents an exponential function with the natural constant as the base. To prevent calculation overflow, the standard unit price standard deviation is used. Scale and reference project volume When the value is 0, the system assigns it a preset minimum positive constant. This ensures that the calculation remains legally closed-loop even under any extreme engineering cost data. In this embodiment, The value is set to 0.001. In other embodiments, implementers can adjust this value according to the actual implementation situation. .

[0032] When the comprehensive unit price Average of standard unit price The larger the absolute value of the difference, and the larger the standard deviation of the standard unit price. The smaller the size, the greater the deviation of the base price from the standard distribution; simultaneously, when the project volume... Compared to the scale of reference engineering volume When it is larger, The smaller the value, the more exponential the function. The closer the value is to 0, the more... When the price approaches its maximum value of 1, it releases the impact of price fluctuations; conversely, if the project volume... Extremely small, even if the price deviates significantly. It will also approach 0, causing the price-volume disturbance energy to... It is suppressed.

[0033] S4. Adjust the price disturbance energy based on the difference between the preset text tolerance boundary parameter and the text variability to obtain the anomaly index of the target cost item.

[0034] It should be noted that, since significant price changes in normal cost logic must be accompanied by modifications to process features of equal depth, it is impossible to accurately determine whether there is a hidden price adjustment behavior based solely on a single anomaly in price or text. Therefore, this invention couples the textual variability of the target cost item with the price-quantity disturbance energy to construct an anomaly index. Through cross-dimensional collaborative calculation, cost items that do not conform to the inherent correlation logic are extracted.

[0035] Specifically, the anomaly index of the target cost item satisfies the expression: In the formula, Indicates the first Abnormal index of each target cost item; Indicates the first The price-quantity disturbance energy of each target cost item; Indicates the first Textual variability of each target cost item; Represents the maximum value function; To prevent extremely small positive constants with a denominator of 0, and to prevent calculation overflow errors caused by division by 0, this embodiment will... Set to the minimum value of text variability greater than zero among all historical cost entries. In other embodiments, the implementer may perform the calculation based on the limitations of the system's computational accuracy. Settings; The tolerance boundary for text variation is determined by: acquiring all historical normal cost items that have undergone reasonable technological changes, statistically analyzing the text variation of these items, obtaining the distribution interval of text variation, and using the lower limit of this distribution interval as the text tolerance boundary parameter. ; This represents the hyperbolic tangent function, used to adjust the difference. Perform nonlinear mapping operations.

[0036] When text variability Minimal energy of price-quantity perturbation When the value is maximized, it indicates that the unit price of the target cost item has changed significantly without substantial changes to the feature text. At this point, the text tolerance boundary parameter... With text variability The difference is a large positive value. Approaching 1, making Close to 2, and in the maximum function The combined effect of the smallest denominator in the output leads to the abnormal index. A non-linear increase in text variability enables anomaly localization; conversely, a decrease in text variability... A sufficiently large value indicates that the change in unit price is supported by semantic alteration, thus defining the text tolerance boundary parameter. With text variability The difference is negative. A negative value indicates a significant decrease in the anomaly index. .

[0037] S5. In response to an abnormal index exceeding the warning threshold, mark the target cost item as a high-risk item and output an intelligent quality control diagnostic report.

[0038] Specifically, the abnormality index of all target cost items is extracted. If the abnormality index of any target cost item exceeds the warning threshold, that target cost item is marked as a high-risk item with pricing defects. The warning threshold is set as follows: The abnormality index of normal cost items that have been manually reviewed and determined to be free of pricing defects is obtained. These abnormality indices are statistically analyzed, and the upper limit of the abnormality index distribution range is obtained as a basic reference value. This basic reference value is multiplied by a preset risk tolerance coefficient and an upward offset is calculated. The result is used as the warning threshold. In this embodiment, the risk tolerance coefficient is 1.2, and the warning threshold is 4.5. In other embodiments, the implementers can set the warning threshold according to the stringency requirements of the specific audit project and the tolerance threshold for underreporting.

[0039] Extract the project feature text, comprehensive unit price, quantity, list classification code, text variability, and price-quantity disturbance energy of all high-risk items, summarize and splice them to generate an intelligent quality control diagnostic report for engineering cost documents and output it.

[0040] For example, Figure 2 This is a schematic diagram of the abnormal index distribution in an embodiment of the present invention. Figure 2 The horizontal axis represents the item number of the target cost, and the vertical axis represents the anomaly index. Figure 2The diagram contains a horizontal dashed line representing the warning threshold. The scatter points for normal cost items and reasonable change items are stably distributed below the warning threshold, while the scatter points for hidden unbalanced pricing items show a vertically rising distribution characteristic, with values ​​greater than the warning threshold. This intuitively demonstrates that the present invention can maintain a stable and extremely low baseline distribution when processing multi-dimensional cost items, and produces a significant nonlinear surge characteristic when encountering abnormally altered items, thus realizing the extraction of high-risk abnormal data.

[0041] Figure 3 This is a schematic diagram illustrating the mapping relationship between text variability and anomaly index in an embodiment of the present invention. Figure 3 The horizontal axis represents text variability, and the vertical axis represents the anomaly index. Figure 3 The system includes a vertical reference line representing the text tolerance boundary parameter and a horizontal reference line representing the warning threshold. When the text variability of a cost item is extremely small and close to the vertical axis, its corresponding anomaly index rises sharply along the vertical axis and exceeds the warning threshold, reflecting the system's interception characteristic of unit price changes without text changes. However, when the text variability of a cost item exceeds the text tolerance boundary parameter, regardless of the magnitude of its price and quantity disturbance energy, the anomaly index is affected by exponential decay and falls below the warning threshold.

[0042] Figure 4 This is a schematic diagram comparing the unit price deviation of the traditional method with the anomaly index of this invention. Figure 4 The horizontal axis represents the deviation from the traditional unit price, and the vertical axis represents the anomaly index. Figure 4 The system includes a vertical dashed line representing the traditional static alarm line and a horizontal dashed line representing the warning threshold. Within the conventional judgment area to the left of the traditional static alarm line, the anomaly index of hidden unbalanced pricing items jumps and exceeds the horizontal warning threshold, effectively identifying anomaly items hidden by minute deviations and compensating for the underreporting defects of traditional methods. In the over-limit judgment area to the right of the traditional static alarm line, the anomaly index of cost items with reasonable changes and minimal engineering quantities is calculated jointly by the engineering quantity attenuation and text tolerance, and is less than the warning threshold, effectively eliminating the false alarm phenomenon of traditional mechanisms. This verifies the ability of the multi-dimensional mapping relationship constructed in this invention to identify hidden defects in engineering cost.

[0043] This invention also discloses an intelligent quality control system for engineering cost documents based on multi-dimensional data, including a processor and a memory. The memory stores computer program instructions, and when the computer program instructions are executed by the processor, the intelligent quality control method for engineering cost documents based on multi-dimensional data according to this invention is implemented.

[0044] The system also includes other components well known to those skilled in the art, such as communication buses and communication interfaces, the settings and functions of which are known in the art and will not be described in detail here.

[0045] In the description of this specification, "multiple" or "several" means at least two, such as two, three or more, unless otherwise expressly and specifically defined.

[0046] While this specification has shown and described numerous embodiments of the invention, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Many modifications, alterations, and alternatives will occur to those skilled in the art without departing from the spirit and essence of the invention. It should be understood that various alternatives to the embodiments of the invention described herein may be employed in the practice of this invention.

Claims

1. An intelligent quality control method for engineering cost documents based on multi-dimensional data, characterized in that, include: Parse the engineering cost document to obtain the target cost item, and extract the project feature text, comprehensive unit price, quantity and list classification code of the target cost item; Based on the list classification code, retrieve matching standard reference items from the historical benchmark database, and extract the standard reference text, standard unit price distribution data, and scale reference quantities corresponding to the standard reference items; Extract core process feature words from the project feature text and standard reference text. Based on the number of missing core process feature words and the string edit distance between the project feature text and the standard reference text, determine the text variability of the target cost item. The deviation of the comprehensive unit price from the standard unit price distribution data is calculated, and the deviation is adjusted using the ratio of the quantity of work to the scale reference quantity of work to obtain the price-quantity disturbance energy of the target cost item. The price disturbance energy is adjusted based on the difference between the preset text tolerance boundary parameter and the text variability to obtain the anomaly index of the target cost item. In response to an abnormal index exceeding the warning threshold, the target cost item is marked as a high-risk item and an intelligent quality control diagnostic report is output.

2. The intelligent quality control method for engineering cost documents based on multi-dimensional data according to claim 1, characterized in that, The step of retrieving matching standard reference entries from the historical benchmark database based on the list classification code includes: Based on the list classification code, retrieve from the historical benchmark database the item that belongs to the same category as the target cost item and has the smallest difference in overall project volume, and use it as the standard reference item.

3. The intelligent quality control method for engineering cost documents based on multi-dimensional data according to claim 1, characterized in that, The extracted core process feature words from the project feature text and standard reference text include: By using word segmentation algorithms and engineering-specific dictionaries, the project feature text and standard reference text are segmented and parsed to extract several core process feature words contained in the project feature text and standard reference text.

4. The intelligent quality control method for engineering cost documents based on multi-dimensional data according to claim 1 or 3, characterized in that, The determination of textual variability of the target cost item includes: The number of core process feature words in the standard reference text that are present but not covered by the core process feature words of the target cost item is counted to obtain the number of missing feature words; the edit distance between the project feature text and the standard reference text is calculated, and the maximum character length of the project feature text and the standard reference text is obtained. The edit distance is normalized using the maximum character length to obtain the distance deviation; a feature missing penalty term is constructed based on the ratio between the number of missing feature words and the total number of core process feature words in the standard reference text; and the text variability is determined by combining the feature missing penalty term and the distance deviation.

5. The intelligent quality control method for engineering cost documents based on multi-dimensional data according to claim 1, characterized in that, The calculation of the deviation of the comprehensive unit price from the standard unit price distribution data, and the adjustment of the deviation using the ratio of the quantity of work to the scale reference quantity of work, yields the price-quantity disturbance energy of the target cost item, including: Extract the standard unit price mean and standard unit price standard deviation from the standard unit price distribution data, calculate the difference between the comprehensive unit price and the standard unit price mean, extract the absolute value of the calculation result and divide it by the standard unit price standard deviation to obtain the basic price deviation; perform natural exponential mapping on the negative number of the ratio of the quantity of work to the scale reference quantity of work to obtain the attenuation median value, and use the difference between the value 1 and the attenuation median value as the quantity of work release coefficient; multiply the basic price deviation by the quantity of work release coefficient to obtain the price and quantity disturbance energy.

6. The intelligent quality control method for engineering cost documents based on multi-dimensional data according to claim 1, characterized in that, The difference between the preset text tolerance boundary parameter and the text variability is used to adjust the price quantity perturbation energy to obtain the anomaly index of the target cost item, including: The maximum value between the text variability and the preset minimum positive constant is obtained as the anti-overflow denominator; the basic disturbance term is determined based on the ratio of the price-quantity disturbance energy to the anti-overflow denominator; a nonlinear dynamic damping term is constructed using the hyperbolic tangent function based on the difference between the text tolerance boundary parameter and the text variability; the nonlinear dynamic damping term is used to adjust the basic disturbance term through product to obtain the anomaly index of the target cost item.

7. The intelligent quality control method for engineering cost documents based on multi-dimensional data according to claim 1, characterized in that, The output intelligent quality control diagnostic report includes: Extract the project feature text, comprehensive unit price, quantity, list classification code, text variability, and price-quantity disturbance energy of all high-risk items, summarize and splice them to generate an intelligent quality control diagnostic report for engineering cost documents and output it.

8. An intelligent quality control system for engineering cost documents based on multi-dimensional data, characterized in that: include: A processor and a memory, wherein the memory stores computer program instructions that, when executed by the processor, implement the intelligent quality control method for engineering cost documents based on multi-dimensional data according to any one of claims 1-7.

Images