Innovative project evaluation method, device, computer equipment and readable storage medium
By constructing a knowledge graph of innovation projects, collecting multi-source data, and adjusting scores, the problems of subjectivity and unfairness in traditional evaluation methods are solved, and fair and reliable scoring and resource allocation of innovation projects are achieved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- GUANGZHOU POWER SUPPLY BUREAU GUANGDONG POWER GRID CO LTD
- Filing Date
- 2026-02-11
- Publication Date
- 2026-06-05
AI Technical Summary
Traditional evaluation methods for innovation projects are highly subjective and lack bias correction and fairness calibration mechanisms, resulting in inaccurate and unfair evaluation results that affect the rationality of decision-making and resource allocation.
By constructing a knowledge graph of innovative projects, collecting multi-source data, estimating the processing bias score and expected project results, calculating the biased raw score, and adjusting the score with the goal of limiting the score differences between groups, an interpretable evaluation report is finally generated.
It improves the fairness and transparency of evaluation results, reduces resource waste, ensures the authenticity and reliability of scores, and supports rational resource allocation and decision-making efficiency.
Smart Images
Figure CN122155492A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the fields of innovation governance and artificial intelligence technology, and in particular to an innovation project evaluation method, apparatus, computer equipment, computer-readable storage medium, and computer program product. Background Technology
[0002] As the innovation environment changes, the evaluation of innovation projects has gradually become a key indicator for measuring the innovation capabilities of enterprises and organizations.
[0003] Traditional methods for evaluating innovation projects mainly include the Analytic Hierarchy Process (AHP), Data Envelopment Analysis (DEA), and Principal Component Analysis (PCA). AHP constructs a multi-level decision-making structure and relies on expert judgment to determine the importance of each factor, providing support for innovation project evaluation decisions. DEA is primarily used to assess the relative efficiency of decision-making units, and can be conducted without pre-setting weights, making it suitable for multi-input, multi-output scenarios. PCA, as a dimensionality reduction technique, effectively reduces redundant features and improves computational efficiency by transforming the original data into a small number of principal components.
[0004] However, these methods generally suffer from high subjectivity, especially the analytic hierarchy process (AHP), which relies excessively on expert judgment and is susceptible to personal bias. Furthermore, they lack effective bias correction mechanisms, meaning that in complex innovation resource environments, evaluation results may be influenced by external factors, causing scores to fail to reflect true innovation capabilities. In addition, while data envelopment analysis (DEA) and principal component analysis (PCA) can reduce subjective bias to some extent, they still lack effective standardization when comparing performance across domains and projects, making scores from different projects incomparable and affecting the fairness and rationality of decision-making.
[0005] Therefore, how to construct an evaluation method for innovative projects that can provide a mechanism for debiasing and fair calibration is an urgent problem to be solved. Summary of the Invention
[0006] Therefore, it is necessary to provide an innovative project evaluation method, apparatus, computer equipment, computer-readable storage medium, and computer program product that can provide a debiasing and fair calibration mechanism to address the above-mentioned technical problems.
[0007] Firstly, this application provides a method for evaluating innovative projects, including:
[0008] Collect data on innovative projects from multiple sources;
[0009] Based on innovation project data, construct an innovation project knowledge graph;
[0010] Based on the knowledge graph of innovation projects, estimate the processing tendency score and expected outcome of at least one innovation project under the given preset background features; calculate the bias-free original score of at least one innovation project based on the processing tendency score and expected outcome.
[0011] With the goal of limiting the score differences of at least one innovative project among different groups, the biased original scores are adjusted to obtain the innovative project score of at least one innovative project.
[0012] The innovation project scores are correlated with the innovation project knowledge graph to generate an innovation project evaluation report; the innovation project evaluation report is used to explain the innovation project scores.
[0013] In one embodiment, an innovation project knowledge graph is constructed based on innovation project data, including:
[0014] Extract multiple entities from the innovation project data, as well as the relationships between these entities;
[0015] A knowledge graph for innovation projects is constructed based on multiple entities and the relationships between them.
[0016] In one embodiment, based on an innovation project knowledge graph, estimating the processing tendency and expected outcome of at least one innovation project given preset background features includes:
[0017] Based on the knowledge graph of innovation projects, obtain the feature vector of at least one innovation project;
[0018] Given preset background features, estimate the processing tendency score and expected outcome of at least one innovative project based on feature vectors.
[0019] In one embodiment, based on the processing propensity score and the expected outcome of the project, a debiased raw score for at least one innovative project is calculated, including:
[0020] Based on the propensity score and the expected project outcome, calculate the objective debiasing score for at least one innovative project;
[0021] The estimation accuracy of the target debiasing score is evaluated. If the estimation accuracy does not meet the first preset condition, the process of estimating the processing tendency score and expected result of at least one innovative project based on the innovative project knowledge graph and given preset background features, as well as subsequent steps, is repeated until the estimation accuracy meets the first preset condition. The corresponding target debiasing score is then used as the original debiasing score.
[0022] In one embodiment, with the goal of limiting the score differences of at least one innovative project among different groups, the biased original score is adjusted to obtain the innovation project score of at least one innovative project, including:
[0023] Obtain group variables; based on the group variables, identify multiple groups to which at least one innovation project belongs;
[0024] Based on the biased original scores, the score differences between multiple groups are calculated.
[0025] If the score difference does not meet the second preset condition, iteratively adjust the original score to remove bias until the score difference meets the second preset condition, and obtain the innovation project score of at least one innovation project.
[0026] In one embodiment, an association analysis is performed between the innovation project score and the innovation project knowledge graph to generate an innovation project evaluation report, including:
[0027] By performing correlation analysis between innovation project scores and innovation project knowledge graphs, a chain of evidence for scoring decisions is generated for innovation project scores.
[0028] Based on the scoring decision evidence chain, an evaluation report for innovative projects is generated.
[0029] Secondly, this application also provides an innovative project evaluation device, comprising:
[0030] The data acquisition module is used to collect data from innovative projects from multiple sources;
[0031] The knowledge graph construction module is used to build knowledge graphs for innovation projects based on innovation project data.
[0032] The causal inference and bias-reduction estimation module is used to estimate the treatment tendency score and expected outcome of at least one innovative project based on the innovative project knowledge graph and given preset background features; and to calculate the bias-reduction raw score of at least one innovative project based on the treatment tendency score and expected outcome.
[0033] The group fairness constraint module is used to adjust the biased original score to obtain the innovation project score of at least one innovation project with the goal of limiting the score difference of at least one innovation project among different groups.
[0034] The interpretability and traceability output module is used to perform correlation analysis between innovation project scores and innovation project knowledge graphs to generate innovation project evaluation reports; the innovation project evaluation reports are used to interpret innovation project scores.
[0035] Thirdly, this application also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to perform the following steps:
[0036] Collect data on innovative projects from multiple sources;
[0037] Based on innovation project data, construct an innovation project knowledge graph;
[0038] Based on the knowledge graph of innovation projects, estimate the processing tendency score and expected outcome of at least one innovation project under the given preset background features; calculate the bias-free original score of at least one innovation project based on the processing tendency score and expected outcome.
[0039] With the goal of limiting the score differences of at least one innovative project among different groups, the biased original scores are adjusted to obtain the innovative project score of at least one innovative project.
[0040] The innovation project scores are correlated with the innovation project knowledge graph to generate an innovation project evaluation report; the innovation project evaluation report is used to explain the innovation project scores.
[0041] Fourthly, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, performs the following steps:
[0042] Collect data on innovative projects from multiple sources;
[0043] Based on innovation project data, construct an innovation project knowledge graph;
[0044] Based on the knowledge graph of innovation projects, estimate the processing tendency score and expected outcome of at least one innovation project under the given preset background features; calculate the bias-free original score of at least one innovation project based on the processing tendency score and expected outcome.
[0045] With the goal of limiting the score differences of at least one innovative project among different groups, the biased original scores are adjusted to obtain the innovative project score of at least one innovative project.
[0046] The innovation project scores are correlated with the innovation project knowledge graph to generate an innovation project evaluation report; the innovation project evaluation report is used to explain the innovation project scores.
[0047] Fifthly, this application also provides a computer program product, including a computer program that, when executed by a processor, performs the following steps:
[0048] Collect data on innovative projects from multiple sources;
[0049] Based on innovation project data, construct an innovation project knowledge graph;
[0050] Based on the knowledge graph of innovation projects, estimate the processing tendency score and expected outcome of at least one innovation project under the given preset background features; calculate the bias-free original score of at least one innovation project based on the processing tendency score and expected outcome.
[0051] With the goal of limiting the score differences of at least one innovative project among different groups, the biased original scores are adjusted to obtain the innovative project score of at least one innovative project.
[0052] The innovation project scores are correlated with the innovation project knowledge graph to generate an innovation project evaluation report; the innovation project evaluation report is used to explain the innovation project scores.
[0053] The aforementioned innovation project evaluation methods, devices, computer equipment, computer-readable storage media, and computer program products first collect innovation project data from multiple sources. Based on this data, an innovation project knowledge graph is constructed. This knowledge graph not only provides static information about the innovation projects but also, through dynamic connections of entity relationships, demonstrates the interrelationships and influences between projects, providing a structured foundation for subsequent data analysis and reasoning. Second, based on the knowledge graph, and given preset background features, the processing tendency score and expected outcome of at least one innovation project are estimated. Based on these scores, a bias-free original score for at least one innovation project is calculated. With the goal of limiting the score differences for at least one innovation project among different groups, the bias-free original score is adjusted to obtain the innovation project score for at least one innovation project. Through this bias-free estimation and fairness constraint process, evaluation bias caused by uneven resource distribution or historical differences is eliminated, improving the fairness and transparency of the evaluation results, reducing resource waste caused by unfair scoring, and making the obtained innovation project scores more realistic and reliable. This contributes to the rational allocation of resources for innovation projects, avoids resource misallocation, and improves innovation efficiency and decision-making efficiency. Finally, by performing correlation analysis between innovation project scores and innovation project knowledge graphs, an innovation project evaluation report is generated. This report provides an explanation of the innovation project scores, supports the traceability of innovation projects, and provides a basis for subsequent project optimization. Attached Figure Description
[0054] To more clearly illustrate the technical solutions in the embodiments of this application or related technologies, the drawings used in the description of the embodiments of this application or related technologies will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.
[0055] Figure 1This is a diagram illustrating the application environment of an innovative project evaluation method in one embodiment.
[0056] Figure 2 This is a flowchart illustrating an innovative project evaluation method in one embodiment;
[0057] Figure 3 This is a flowchart illustrating the causal inference step in an innovative project evaluation method in one embodiment;
[0058] Figure 4 This is a flowchart illustrating the innovation project evaluation method in another embodiment;
[0059] Figure 5 This is a structural block diagram of an innovative project evaluation device in one embodiment;
[0060] Figure 6 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation
[0061] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.
[0062] It should be noted that the terms "first," "second," etc., used in this application can be used to describe various elements, but these elements are not limited by these terms. These terms are only used to distinguish the first element from the second element. The terms "comprising" and "having," and any variations thereof, used in this application, are intended to cover non-exclusive inclusion. The term "multiple" used in this application refers to two or more. The term "and / or" used in this application refers to one of the embodiments, or any combination of multiple embodiments.
[0063] The innovative project evaluation method provided in this application can be applied to, for example... Figure 1In the application environment shown, terminal 102 communicates with server 104 via a network. A data storage system can store the data that server 104 needs to process. The data storage system can be integrated onto server 104 or located in the cloud or on other network servers. Specifically, terminal 102 or server 104 executes an innovation project evaluation method, which includes: collecting innovation project data from multiple sources; constructing an innovation project knowledge graph based on the innovation project data; estimating the processing tendency score and expected outcome of at least one innovation project based on the innovation project knowledge graph, given preset background features; calculating the biased original score of at least one innovation project based on the processing tendency score and expected outcome; adjusting the biased original score to limit the score difference of at least one innovation project among different groups, obtaining an innovation project score for at least one innovation project; performing a correlation analysis between the innovation project score and the innovation project knowledge graph to generate an innovation project evaluation report; and using the innovation project evaluation report to interpret the innovation project score.
[0064] Terminal 102 can be, but is not limited to, various personal computers, laptops, smartphones, tablets, drones, low-altitude aircraft, IoT devices, and portable wearable devices. IoT devices can include smart speakers, smart TVs, smart air conditioners, smart in-vehicle devices, and projection equipment. Portable wearable devices can include smartwatches, smart bracelets, and head-mounted displays. Head-mounted displays can be virtual reality (VR) devices, augmented reality (AR) devices, and smart glasses. Server 104 can be a standalone physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing cloud computing services.
[0065] In one exemplary embodiment, such as Figure 2 As shown, an innovative project evaluation method is provided, which is then applied to... Figure 1 Taking terminal 102 as an example, the explanation includes the following steps 202 to 210. Wherein:
[0066] Step 202: Collect innovation project data from multiple sources.
[0067] The data on innovation projects includes information on innovation input, project output, team size, platform support, and historical accumulation. Innovation input refers to the various resources an organization invests in carrying out innovation project activities, such as funding, resources, and R&D time. Project output refers to the products or services directly created by innovation project activities, such as patents, papers, products, and market impact.
[0068] Step 204: Construct an innovation project knowledge graph based on innovation project data.
[0069] The knowledge graph of innovation projects includes nodes and edges. Nodes represent entities that can be extracted from innovation project data, and edges represent relationships between entities that can be extracted from innovation project data.
[0070] For example, innovation project data is integrated, and entities and relationships between entities are extracted from the innovation project data; based on the entities and relationships between entities, an innovation project knowledge graph is constructed.
[0071] Step 206: Based on the knowledge graph of innovation projects, estimate the processing tendency score and expected outcome of at least one innovation project under the given preset background features; calculate the debiased original score of at least one innovation project based on the processing tendency score and expected outcome.
[0072] The processing propensity score refers to the probability that an innovative project will receive a certain processing given background features; processing refers to a specific operation performed on the innovative project, such as financial support. The debiased raw score must meet a preset accuracy standard.
[0073] For example, feature vectors of at least one innovation project are obtained based on an innovation project knowledge graph. Based on the feature vectors and given preset background features, a propensity score matching method is used to estimate the processing propensity score for each project; and an outcome model is used to calculate the expected project outcome for at least one innovation project.
[0074] Specifically, propensity score matching is used to estimate the treatment propensity for each item, ensuring comparisons are made only when the background conditions of the items are similar. The resulting model is used to characterize the expected outcome of the items given the treatment state and background features.
[0075] For example, based on the processing propensity score and the expected project result, the "double robust" estimation method is used to estimate the debiased original score of at least one innovative project. The "double robust" estimation method can comprehensively process the propensity score and the expected project result, and provide a relatively consistent and unbiased estimate provided that either the processing propensity score or the expected project result is correct, and calculate the debiased original score.
[0076] Step 208: With the goal of limiting the score differences of at least one innovative project among different groups, adjust the biased original score to obtain the innovative project score of at least one innovative project.
[0077] For example, a group refers to a collection of innovative projects after classifying, stratifying, or grouping at least one innovative project using group variables.
[0078] For example, based on group variables, after identifying multiple groups to which at least one innovative project belongs, at least one group is obtained, and each group includes at least one innovative project. An innovative project may belong to multiple different groups. Then, the sum of the debiased raw scores of all innovative projects in any one group is calculated. Based on this sum, the average debiased raw scores of the two groups are calculated. The difference between these average debiased raw scores of any two groups is calculated, and the absolute value of this difference is taken as the score difference between the two groups. Then, with the goal of limiting the score difference of at least one innovative project between different groups, an optimization algorithm is used to adjust the debiased raw scores, and the adjusted debiased raw scores are taken as the scores of the innovative projects.
[0079] Step 210: Perform correlation analysis between the innovation project score and the innovation project knowledge graph to generate an innovation project evaluation report; the innovation project evaluation report is used to explain the innovation project score.
[0080] For example, by associating the score of each innovation project with nodes and edges in the innovation project knowledge graph, it is possible to trace the background information of the innovation project, such as the specific impact of factors like investment, team size, and platform support on the score. Based on the traced results, an innovation project evaluation report is generated.
[0081] Optionally, when new data is received, the knowledge graph structure and the parameters of all computational models can be automatically updated.
[0082] The aforementioned innovation project evaluation method first collects innovation project data from multiple sources. Based on this data, an innovation project knowledge graph is constructed. This knowledge graph not only provides static information about the innovation projects but also, through dynamic connections of entity relationships, demonstrates the interrelationships and influences between projects, providing a structured foundation for subsequent data analysis and reasoning. Second, based on the innovation project knowledge graph, and given preset background features, the processing tendency score and expected outcome of at least one innovation project are estimated. Based on these scores, a bias-free original score for at least one innovation project is calculated. To limit the score differences between different groups for at least one innovation project, the bias-free original score is adjusted to obtain the innovation project score for at least one innovation project. Through this bias-free estimation and fairness constraint process, evaluation bias caused by uneven resource allocation or historical differences is eliminated, improving the fairness and transparency of the evaluation results, reducing resource waste caused by unfair scoring, and making the obtained innovation project scores more realistic and reliable. This contributes to the rational allocation of resources for innovation projects, avoids resource misallocation, and improves innovation efficiency and decision-making efficiency. Finally, by performing correlation analysis between innovation project scores and innovation project knowledge graphs, an innovation project evaluation report is generated. This report provides an explanation of the innovation project scores, supports the traceability of innovation projects, and provides a basis for subsequent project optimization.
[0083] In one exemplary embodiment, constructing an innovation project knowledge graph based on innovation project data includes: extracting multiple entities from the innovation project data and the relationships between the multiple entities; and constructing the innovation project knowledge graph based on the multiple entities and the relationships between the multiple entities.
[0084] The entities include innovative projects, teams, and platforms; the relationships between multiple entities include belonging, influence, and support.
[0085] For example, suppose there is One innovative project, denoted as At this point, each innovation project in the innovation project knowledge graph Corresponding to a node The relationship between innovative projects is achieved through the edge The knowledge graph of innovative projects is represented by the following formula:
[0086]
[0087] in, It is a set of nodes, representing entities such as projects, teams, and resources. It is a set of edges, representing the relationships between these entities. Each edge... Connecting nodes and This indicates the relationship between the two nodes.
[0088] In this embodiment, by extracting multiple entities from the innovation project data and the relationships between these entities, an innovation project knowledge graph is constructed. This graph provides static information about the innovation projects for subsequent data analysis and reasoning. Furthermore, through the dynamic connections between the entities, it demonstrates the interrelationships and influences between the various innovation projects.
[0089] In one embodiment, such as Figure 3 As shown, based on the knowledge graph of innovation projects, the processing tendency and expected outcome of at least one innovation project are estimated under given preset background features, including steps 302 to 304. Wherein:
[0090] Step 302: Obtain the feature vector of at least one innovation project based on the innovation project knowledge graph.
[0091] For example, based on the innovation project knowledge graph, feature vectors for at least one innovation project are obtained. The feature vectors are represented as follows:
[0092]
[0093] in, Number the innovation projects. For innovative projects eigenvectors; This is a mapping between the knowledge graph of innovation projects and the features generated from innovation project data. These are parameters, such as: embedding dimension, time decay coefficient, and aggregation caliber.
[0094] Step 304: Given preset background features, estimate the processing tendency score and expected outcome of at least one innovative project based on the feature vector.
[0095] For example, given preset background features, a propensity score matching method is used to estimate the processing propensity score for at least one innovative item based on feature vectors. The processing propensity score is typically estimated using a logistic regression model, with the following formula:
[0096]
[0097] in, For innovative projects In the eigenvector Probability estimate of acceptance; To process instructions, such as whether to obtain funding or platform support; and For parameter vectors.
[0098] Optionally, the expected project outcome can be calculated using an outcome model under a given processing state and background characteristics, as shown in the following model:
[0099]
[0100] in, Given a feature vector Processing status Innovation projects under conditions The result is predicted. For innovative projects The observation results, such as the overall output score or the overall benefit score. Result prediction can be described as a mathematical expectation.
[0101] In this embodiment, by obtaining the feature vectors of innovative projects and estimating the processing tendency score and expected project outcome based on these vectors under preset background features, the potential performance and likelihood of acceptance of innovative projects in a real environment can be more accurately reflected. The application of the logistic regression model makes the estimation of the processing tendency score statistically rigorous, while the outcome model provides a quantitative prediction of the expected project outcome. The combination of the two lays a solid foundation for the subsequent debiased raw score calculation.
[0102] In one embodiment, calculating the original debiased score of at least one innovation project based on the processing tendency score and the expected project result includes: calculating the target debiased score of at least one innovation project based on the tendency score and the expected project result; evaluating the estimation accuracy of the target debiased score; if the estimation accuracy does not meet a first preset condition, repeatedly performing the steps of estimating the processing tendency score and the expected project result of at least one innovation project based on the innovation project knowledge graph and given preset background features, and subsequent steps, until the estimation accuracy meets the first preset condition, and using the corresponding target debiased score as the original debiased score.
[0103] The first preset condition refers to the estimation accuracy meeting the preset accuracy standard.
[0104] Optionally, the formula for the target debiasing score is as follows:
[0105]
[0106] in, For innovative projects The original score was de-biased. The expected outcome of the project is indicated by the background characteristics. and handling variables The expected results For innovative projects The processing tendency score For innovative projects The actual result It is an indicator function used to indicate innovative projects. Whether to accept the treatment, such as whether to receive financial support.
[0107] In this embodiment, by calculating the target debiased score for at least one innovative project based on the propensity score and the expected project outcome, confounding factors affecting performance can be effectively eliminated, enabling a fair comparison of different innovative projects under similar conditions. The estimation accuracy of the target debiased score is evaluated. If the estimation accuracy does not meet a first preset condition, the process of estimating the processing propensity score and expected project outcome for at least one innovative project based on the innovative project knowledge graph and given preset background features, along with subsequent steps, is repeated until the estimation accuracy meets the first preset condition. The corresponding target debiased score is then used as the original debiased score. This iterative process ensures that the final obtained original debiased score has higher accuracy.
[0108] In one embodiment, with the goal of limiting the score difference of at least one innovative project among different groups, adjusting the biased original score to obtain the innovative project score of at least one innovative project includes: obtaining group variables; determining multiple groups to which at least one innovative project belongs based on the group variables; calculating the score difference between the multiple groups based on the biased original score; and iteratively adjusting the biased original score until the score difference meets the second preset condition if the score difference does not meet the second preset condition, thereby obtaining the innovative project score of at least one innovative project.
[0109] Among these, group variables refer to variables that can classify, stratify, or group at least one innovative project according to different dimensions, such as project type, platform level, etc. The second pre-set condition is a fairness threshold where the score difference is less than or equal to a pre-set value. This pre-set fairness threshold is used to ensure that the score differences between groups remain within a reasonable range.
[0110] For example, the biased original score is iteratively adjusted using the following formula:
[0111]
[0112] in, This indicates the search for the option that minimizes the objective. The optimal result is found Recorded as This refers to the scoring of innovative projects. To remove bias from the original score.
[0113] Optionally, the second preset condition can be expressed as:
[0114]
[0115] in, For group variables, and They represent in the group and The average rating in the middle, The fairness threshold is a preset value.
[0116] In this embodiment, the score difference between multiple groups is calculated based on the original score after bias removal. If the score difference does not meet the second preset condition, the original score after bias removal is iteratively adjusted until the score difference meets the second preset condition, so as to obtain the innovation project score of at least one innovation project. This ensures that the innovation projects can be compared fairly among different groups and avoids the score deviation caused by factors such as resource background and platform support.
[0117] In one embodiment, an innovation project evaluation report is generated by performing a correlation analysis between the innovation project score and the innovation project knowledge graph. This includes: performing a correlation analysis between the innovation project score and the innovation project knowledge graph to generate a scoring decision evidence chain for the innovation project score; and generating an innovation project evaluation report based on the scoring decision evidence chain.
[0118] Optionally, the formula for linking the scoring results with the evidence nodes in the innovation project knowledge graph is as follows:
[0119]
[0120] in, For innovative projects The final score, In knowledge graph In and project nodes The relevant first The contribution of each piece of evidence to the score (carrying source and confidence level); The attribution weights of the corresponding evidence satisfy the normalization constraints of non-negativity and a sum of 1. For innovative projects The upper limit on the number of pieces of evidence that can be included in the interpretation.
[0121] In this embodiment, by correlating the innovation project score with the innovation project knowledge graph, a detailed chain of evidence for the scoring decision can be generated. This chain clearly demonstrates the logical basis and key factors behind the scoring. The innovation project evaluation report generated based on this chain of evidence not only includes the final score of the innovation project but also lists in detail the various pieces of evidence supporting the score and their weights, making the evaluation results highly interpretable and traceable. This evaluation method helps project teams deeply understand the reasons behind the scoring, providing strong data support for subsequent project optimization and improvement. Simultaneously, it provides decision-makers with comprehensive and objective project evaluation criteria, contributing to improved decision-making efficiency and accuracy.
[0122] Next reference Figure 4 The innovative project evaluation method of this application will be illustrated by a specific embodiment.
[0123] Step 1: Collect innovation project data from multiple sources, including innovation input, project output, team size, platform support, and historical accumulation.
[0124] Step 2: Integrate all the innovation project data mentioned above to construct an innovation project knowledge graph. In this graph, nodes represent entities such as innovation projects, teams, and platforms, and edges represent the relationships between them.
[0125] Step 3: Based on the knowledge graph of innovation projects, under the given preset background features, use propensity score matching to estimate the treatment propensity score of at least one innovation project, and use the outcome model to estimate the expected results of the project; based on the treatment propensity score and the expected results of the project, use the "double robust" estimation method to estimate the debiased original score of at least one innovation project, which can remove external biases in the data, ensure fair comparison of innovation projects under similar conditions, and thus accurately estimate the causal effect of innovation projects.
[0126] Step 4: Evaluate the accuracy of the causal effect estimation, i.e. the accuracy of the biased original score estimation. If the evaluation is not up to standard, return to step 3 and repeat. If the evaluation is up to standard, continue to step 5.
[0127] Step 5: Apply group fairness constraints to the debiased original scores. This means that the goal is to limit the score differences of at least one innovative project among different groups. By using an optimization algorithm to adjust the debiased original scores, we can ensure that innovative projects are evaluated fairly among different groups.
[0128] Step 6: Perform correlation analysis between the innovation project score and the edges and nodes of the innovation project knowledge graph to generate a scoring decision evidence chain for the innovation project score; based on the scoring decision evidence chain, output an innovation project evaluation report to ensure that each scoring decision can be traced back to a specific source of evidence.
[0129] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages in other steps. It is understood that the steps in different embodiments can be freely combined as needed, and all non-contradictory solutions formed by such combinations are within the scope of protection of this application.
[0130] Based on the same inventive concept, this application also provides an innovation project evaluation device for implementing the innovation project evaluation method described above. The solution provided by this device is similar to the solution described in the above method; therefore, the specific limitations in one or more embodiments of the innovation project evaluation device provided below can be found in the limitations of the innovation project evaluation method described above, and will not be repeated here.
[0131] In one exemplary embodiment, such as Figure 5 As shown, an innovation project evaluation device 500 is provided, including: a data acquisition module 502, a knowledge graph construction module 504, a causal inference and bias-correction estimation module 506, a group fairness constraint module 508, and an interpretability and source tracing output module 510, wherein:
[0132] The data acquisition module 502 is used to collect innovation project data from multiple sources.
[0133] The knowledge graph construction module 504 is used to construct an innovation project knowledge graph based on innovation project data.
[0134] The causal inference and debiasing estimation module 506 is used to estimate the treatment tendency score and expected outcome of at least one innovation project based on the innovation project knowledge graph and given preset background features; and to calculate the debiased original score of at least one innovation project based on the treatment tendency score and expected outcome.
[0135] The group fairness constraint module 508 is used to adjust the biased original score to obtain the innovation project score of at least one innovation project with the goal of limiting the score difference of at least one innovation project among different groups.
[0136] The interpretability and traceability output module 510 is used to perform correlation analysis between the innovation project score and the innovation project knowledge graph to generate an innovation project evaluation report; the innovation project evaluation report is used to interpret the innovation project score.
[0137] In one embodiment, the knowledge graph construction module is also used to extract multiple entities from the innovation project data, as well as the relationships between the multiple entities; and to construct an innovation project knowledge graph based on the multiple entities and the relationships between the multiple entities.
[0138] In one embodiment, the causal inference and debiasing estimation module is further configured to obtain the feature vector of at least one innovative project based on the innovative project knowledge graph; and, given preset background features, estimate the processing tendency score and expected project result of at least one innovative project based on the feature vector.
[0139] In one embodiment, the causal inference and debiasing estimation module is further configured to calculate the target debiasing score of at least one innovation project based on the propensity score and the expected project result; evaluate the estimation accuracy of the target debiasing score; if the estimation accuracy does not meet the first preset condition, repeat the steps of estimating the processing propensity score and the expected project result of at least one innovation project based on the innovation project knowledge graph and given preset background features, and the subsequent steps, until the estimation accuracy meets the first preset condition, and use the corresponding target debiasing score as the original debiasing score.
[0140] In one embodiment, the group fairness constraint module is further used to obtain group variables; based on the group variables, determine multiple groups to which at least one innovation project belongs; based on the debiased original score, calculate the score difference between the multiple groups; if the score difference does not meet the second preset condition, iteratively adjust the debiased original score until the score difference meets the second preset condition, thereby obtaining the innovation project score of at least one innovation project.
[0141] In one embodiment, the interpretability and traceability output module is also used to perform correlation analysis between the innovation project score and the innovation project knowledge graph to generate a scoring decision evidence chain for the innovation project score; and to generate an innovation project evaluation report based on the scoring decision evidence chain.
[0142] Each module in the aforementioned innovation project evaluation device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the operations corresponding to each module.
[0143] In one exemplary embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 6As shown, this computer device includes a processor, memory, input / output interfaces (I / O), and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operating system and computer programs stored in the non-volatile storage media. The database stores innovation project data. The I / O interfaces are used for exchanging information between the processor and external devices. The communication interface is used for communicating with external terminals via a network connection. When executed by the processor, the computer program implements an innovation project evaluation method.
[0144] Those skilled in the art will understand that Figure 6 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.
[0145] In one exemplary embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to perform the following steps:
[0146] Collect data on innovative projects from multiple sources;
[0147] Based on innovation project data, construct an innovation project knowledge graph;
[0148] Based on the knowledge graph of innovation projects, estimate the processing tendency score and expected outcome of at least one innovation project under the given preset background features; calculate the bias-free original score of at least one innovation project based on the processing tendency score and expected outcome.
[0149] With the goal of limiting the score differences of at least one innovative project among different groups, the biased original scores are adjusted to obtain the innovative project score of at least one innovative project.
[0150] The innovation project scores are correlated with the innovation project knowledge graph to generate an innovation project evaluation report; the innovation project evaluation report is used to explain the innovation project scores.
[0151] In one embodiment, when the processor executes the computer program, it also performs the following steps: extracting multiple entities from the innovation project data, as well as the relationships between the multiple entities; and constructing an innovation project knowledge graph based on the multiple entities and the relationships between the multiple entities.
[0152] In one embodiment, when the processor executes the computer program, it further performs the following steps: obtaining feature vectors of at least one innovation project based on the innovation project knowledge graph; and estimating the processing tendency score and expected outcome of at least one innovation project based on the feature vectors, given preset background features.
[0153] In one embodiment, when the processor executes the computer program, it further performs the following steps: calculating the target debiasing score of at least one innovation project based on the propensity score and the expected project result; evaluating the estimation accuracy of the target debiasing score; if the estimation accuracy does not meet a first preset condition, repeatedly executing the steps of estimating the processing propensity score and the expected project result of at least one innovation project based on the innovation project knowledge graph and given preset background features, and subsequent steps, until the estimation accuracy meets the first preset condition, and using the corresponding target debiasing score as the original debiasing score.
[0154] In one embodiment, when the processor executes the computer program, it further performs the following steps: obtaining a group variable; determining multiple groups to which at least one innovation project belongs based on the group variable; calculating the score difference between the multiple groups based on the debiased original score; and iteratively adjusting the debiased original score until the score difference meets the second preset condition if the score difference does not meet the second preset condition, thereby obtaining the innovation project score of at least one innovation project.
[0155] In one embodiment, when the processor executes the computer program, it further performs the following steps: performing correlation analysis between the innovation project score and the innovation project knowledge graph to generate a scoring decision evidence chain for the innovation project score; and generating an innovation project evaluation report based on the scoring decision evidence chain.
[0156] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, the computer program performing the following steps when executed by a processor:
[0157] Collect data on innovative projects from multiple sources;
[0158] Based on innovation project data, construct an innovation project knowledge graph;
[0159] Based on the knowledge graph of innovation projects, estimate the processing tendency score and expected outcome of at least one innovation project under the given preset background features; calculate the bias-free original score of at least one innovation project based on the processing tendency score and expected outcome.
[0160] With the goal of limiting the score differences of at least one innovative project among different groups, the biased original scores are adjusted to obtain the innovative project score of at least one innovative project.
[0161] The innovation project scores are correlated with the innovation project knowledge graph to generate an innovation project evaluation report; the innovation project evaluation report is used to explain the innovation project scores.
[0162] In one embodiment, when the computer program is executed by the processor, it further performs the following steps: extracting multiple entities from the innovation project data, as well as the relationships between the multiple entities; and constructing an innovation project knowledge graph based on the multiple entities and the relationships between the multiple entities.
[0163] In one embodiment, when the computer program is executed by the processor, it further performs the following steps: obtaining feature vectors of at least one innovation project based on the innovation project knowledge graph; and estimating the processing tendency score and expected outcome of at least one innovation project based on the feature vectors, given preset background features.
[0164] In one embodiment, when the computer program is executed by the processor, it further performs the following steps: calculating the target debiasing score of at least one innovation project based on the propensity score and the expected project result; evaluating the estimation accuracy of the target debiasing score; if the estimation accuracy does not meet a first preset condition, repeatedly performing the steps of estimating the processing propensity score and the expected project result of at least one innovation project based on the innovation project knowledge graph and given preset background features, and subsequent steps, until the estimation accuracy meets the first preset condition, and using the corresponding target debiasing score as the original debiasing score.
[0165] In one embodiment, when the computer program is executed by the processor, it further performs the following steps: obtaining group variables; determining multiple groups to which at least one innovation project belongs based on the group variables; calculating the score difference between the multiple groups based on the debiased original score; and iteratively adjusting the debiased original score until the score difference meets the second preset condition if the score difference does not meet the second preset condition, thereby obtaining the innovation project score of at least one innovation project.
[0166] In one embodiment, when the computer program is executed by the processor, it further performs the following steps: performing correlation analysis between the innovation project score and the innovation project knowledge graph to generate a scoring decision evidence chain for the innovation project score; and generating an innovation project evaluation report based on the scoring decision evidence chain.
[0167] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, performs the following steps:
[0168] Collect data on innovative projects from multiple sources;
[0169] Based on innovation project data, construct an innovation project knowledge graph;
[0170] Based on the knowledge graph of innovation projects, estimate the processing tendency score and expected outcome of at least one innovation project under the given preset background features; calculate the bias-free original score of at least one innovation project based on the processing tendency score and expected outcome.
[0171] With the goal of limiting the score differences of at least one innovative project among different groups, the biased original scores are adjusted to obtain the innovative project score of at least one innovative project.
[0172] The innovation project scores are correlated with the innovation project knowledge graph to generate an innovation project evaluation report; the innovation project evaluation report is used to explain the innovation project scores.
[0173] In one embodiment, when the computer program is executed by the processor, it further performs the following steps: extracting multiple entities from the innovation project data, as well as the relationships between the multiple entities; and constructing an innovation project knowledge graph based on the multiple entities and the relationships between the multiple entities.
[0174] In one embodiment, when the computer program is executed by the processor, it further performs the following steps: obtaining feature vectors of at least one innovation project based on the innovation project knowledge graph; and estimating the processing tendency score and expected outcome of at least one innovation project based on the feature vectors, given preset background features.
[0175] In one embodiment, when the computer program is executed by the processor, it further performs the following steps: calculating the target debiasing score of at least one innovation project based on the propensity score and the expected project result; evaluating the estimation accuracy of the target debiasing score; if the estimation accuracy does not meet a first preset condition, repeatedly performing the steps of estimating the processing propensity score and the expected project result of at least one innovation project based on the innovation project knowledge graph and given preset background features, and subsequent steps, until the estimation accuracy meets the first preset condition, and using the corresponding target debiasing score as the original debiasing score.
[0176] In one embodiment, when the computer program is executed by the processor, it further performs the following steps: obtaining group variables; determining multiple groups to which at least one innovation project belongs based on the group variables; calculating the score difference between the multiple groups based on the debiased original score; and iteratively adjusting the debiased original score until the score difference meets the second preset condition if the score difference does not meet the second preset condition, thereby obtaining the innovation project score of at least one innovation project.
[0177] In one embodiment, when the computer program is executed by the processor, it further performs the following steps: performing correlation analysis between the innovation project score and the innovation project knowledge graph to generate a scoring decision evidence chain for the innovation project score; and generating an innovation project evaluation report based on the scoring decision evidence chain.
[0178] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.
[0179] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, artificial intelligence (AI) processors, etc., and are not limited to these.
[0180] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.
[0181] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.
Claims
1. An innovative project evaluation method, characterized in that, The method includes: Collect data on innovative projects from multiple sources; Based on the aforementioned innovation project data, an innovation project knowledge graph is constructed; Based on the knowledge graph of the innovation projects, under the given preset background features, estimate the processing tendency score and expected outcome of at least one innovation project; based on the processing tendency score and expected outcome, calculate the bias-free original score of the at least one innovation project. With the goal of limiting the scoring differences of the at least one innovative project among different groups, the biased original score is adjusted to obtain the innovation project score of the at least one innovative project; The innovation project scores are correlated with the innovation project knowledge graph to generate an innovation project evaluation report; the innovation project evaluation report is used to explain the innovation project scores.
2. The method according to claim 1, characterized in that, The construction of an innovation project knowledge graph based on the innovation project data includes: Extract multiple entities from the innovation project data, as well as the relationships between these entities; Based on the multiple entities and the relationships between them, an innovation project knowledge graph is constructed.
3. The method according to claim 1, characterized in that, The step of estimating the processing tendency and expected outcome of at least one innovative project based on the innovative project knowledge graph, given preset background features, includes: Based on the innovation project knowledge graph, obtain the feature vector of at least one innovation project; Given preset background features, the processing tendency score and expected outcome of the at least one innovative project are estimated based on the feature vector.
4. The method according to claim 1, characterized in that, The calculation of the unbiased raw score for the at least one innovative project based on the processing propensity score and the expected project results includes: Based on the propensity score and the expected project results, calculate the target debiasing score for the at least one innovative project; The estimation accuracy of the target debiasing score is evaluated. If the estimation accuracy does not meet the first preset condition, the process of estimating the processing tendency score and expected result of at least one innovative project based on the innovative project knowledge graph and given preset background features, and the subsequent steps, is repeated until the estimation accuracy meets the first preset condition. The corresponding target debiasing score is then used as the original debiasing score.
5. The method according to claim 1, characterized in that, The step of adjusting the biased original scores to obtain the innovation project scores for the at least one innovative project, with the aim of limiting the score differences of the at least one innovative project among different groups, includes: Obtain group variables; based on the group variables, determine multiple groups to which the at least one innovation project belongs; Based on the original score after bias removal, the score differences among the multiple groups are calculated; If the score difference does not meet the second preset condition, the original score is iteratively adjusted until the score difference meets the second preset condition, thereby obtaining the innovation project score of the at least one innovation project.
6. The method according to claim 1, characterized in that, The step of performing correlation analysis between the innovation project score and the innovation project knowledge graph to generate an innovation project evaluation report includes: The innovation project scores are correlated with the innovation project knowledge graph to generate a chain of evidence for the scoring decision of the innovation project scores; Based on the aforementioned chain of evidence for scoring decisions, an evaluation report for innovative projects is generated.
7. An innovative project evaluation device, characterized in that, The device includes: The data acquisition module is used to collect data from innovative projects from multiple sources; The knowledge graph construction module is used to construct an innovation project knowledge graph based on the innovation project data. The causal inference and bias-reduction estimation module is used to estimate the processing tendency score and expected outcome of at least one innovative project based on the innovative project knowledge graph and given preset background features; and to calculate the bias-reduction original score of the at least one innovative project based on the processing tendency score and expected outcome. The group fairness constraint module is used to adjust the biased original score to obtain the innovation project score of the at least one innovation project with the goal of limiting the score difference of the at least one innovation project among different groups. The interpretability and traceability output module is used to perform correlation analysis between the innovation project score and the innovation project knowledge graph to generate an innovation project evaluation report; the innovation project evaluation report is used to interpret the innovation project score.
8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 6.
9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.
10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.