Automobile innovation project credibility evaluation method and system based on knowledge graph

Abstract

The invention discloses an automobile innovation project credibility evaluation method and system based on a knowledge graph, and relates to the technical field of data processing, and the method comprises the steps: obtaining an implemented automobile project based on the knowledge graph, extracting automobile field keywords, forming an automobile keyword library, obtaining text information, and storing the text information in a database; obtaining an automobile field keyword appearing in the text information and taking the automobile field keyword as a reference word, and obtaining an independent frequency; obtaining a first association degree of every two reference words in the implemented automobile item, and obtaining a first reference credibility; obtaining a time node of each implemented automobile project, obtaining a second correlation degree, and obtaining a second reference credibility corresponding to the automobile innovation project based on the reference model, the second correlation degree and a plurality of independent frequencies; and obtaining a target credibility according to the first reference credibility and the second reference credibility, and associating the target credibility with the automobile innovation project. The method has the advantages of high accuracy, high timeliness, stability and reliability.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and particularly relates to a method and system for evaluating the credibility of automotive innovation projects based on a knowledge graph. Background Art

[0002] With the rapid development of the automotive industry, the number and technical complexity of automotive innovation projects have increased significantly. How to scientifically and efficiently evaluate the technical feasibility and market credibility of projects has become the focus of industry attention. Traditional evaluation methods mainly rely on expert experience or simple analysis based on statistical data, such as predicting through keyword matching or the success rate of historical projects. However, such methods have obvious limitations: on the one hand, expert experience is greatly affected by subjective factors and it is difficult to achieve standardized evaluation; on the other hand, existing data analysis methods lack in-depth exploration of multi-dimensional association relationships within the technical field, especially in complex technical scenarios (such as cross-field technology integration, dynamic technology trend changes), and the accuracy of evaluation results significantly decreases.

[0003] Currently, some studies have attempted to introduce knowledge graph technology to improve the evaluation effect by constructing a knowledge base in the automotive field to reveal the relevance between technical concepts. However, there are still deficiencies in the application of knowledge graphs in existing methods. First, in the calculation of the degree of association, usually only the co-occurrence frequency of keywords is considered, while ignoring the impact of the time evolution characteristics of different historical projects on the strength of technical association. In addition, the processing of the independent frequency of keywords by traditional methods is too simplistic, without considering the noise interference that may be brought by high-frequency words, resulting in poor adaptability of the evaluation model to emerging technologies or sub-domains. For example, certain technology combinations frequently appeared in early projects, but may no longer be applicable with technological iteration, and existing methods do not introduce a time decay mechanism to correct such association relationships; at the same time, in terms of keyword weight assignment, there is a lack of non-linear processing of frequency data, making high-frequency but low-value words have excessive interference on the evaluation results. These problems make it difficult to meet the needs of evaluating the credibility of an increasing number of automotive innovation projects. Summary of the Invention

[0004] Aiming at the defects in the prior art, the present invention provides a method and system for evaluating the credibility of automotive innovation projects based on a knowledge graph.

[0005] A method for evaluating the credibility of an automobile innovation project based on a knowledge graph comprises: obtaining implemented automobile projects based on a knowledge graph, extracting multiple automobile-related keywords from the implemented automobile projects and forming an automobile keyword library, obtaining text information of the automobile innovation project, obtaining automobile-related keywords appearing in the text information as benchmark words based on the text information and the automobile keyword library, and obtaining the number of times each benchmark word appears in the text information as an independent frequency; obtaining a first correlation between every two benchmark words in the implemented automobile projects, and obtaining a first benchmark credibility corresponding to the automobile innovation project based on a benchmark model, the first correlation and multiple independent frequencies; obtaining a time node of each implemented automobile project, and obtaining a second correlation between every two benchmark words in the implemented automobile projects based on the time node of each implemented automobile project, and obtaining a second benchmark credibility corresponding to the automobile innovation project based on the benchmark model, the second correlation and multiple independent frequencies; obtaining a target credibility based on the first benchmark credibility and the second benchmark credibility, and associating the target credibility with the automobile innovation project.

[0006] Optionally, obtaining the first correlation between every two benchmark words in the implemented automobile projects includes: obtaining the number of times every two benchmark words appear at the same time in each implemented automobile project and recording it as the number of correlations; and obtaining the first correlation corresponding to every two benchmark words based on multiple number of correlations corresponding to every two benchmark words.

[0007] Optionally, the first correlation degree corresponding to each two benchmark words is obtained according to a plurality of correlation times corresponding to each two benchmark words, and is expressed as: , ;in, is the first correlation between the i-th benchmark word and the j-th benchmark word, is the number of implemented automotive projects, is the number of occurrences of the i-th benchmark word and the j-th benchmark word in the k-th implemented automotive project.

[0008] Optionally, the benchmark model in the first benchmark credibility corresponding to the automobile innovation project obtained based on the benchmark model, the first correlation and multiple independent frequencies is expressed as: ;in, is the first benchmark credibility, is the first correlation between the i-th benchmark word and the j-th benchmark word, is the number of first degree of association, is the independent frequency corresponding to the Pth benchmark word, is the number of base words.

[0009] Optionally, obtaining the second correlation degree of every two reference words in the implemented automotive projects based on the time nodes of the implemented automotive projects includes: obtaining the number of times of simultaneous occurrence of every two reference words in each implemented automotive project and recording it as the correlation times; obtaining the time increase and decrease weights corresponding to each implemented automotive project according to the time nodes of the implemented automotive projects; and obtaining the second correlation degree corresponding to every two reference words according to the time increase and decrease weights corresponding to each implemented automotive project and the multiple correlation times corresponding to every two reference words.

[0010] Optionally, obtaining the time increase and decrease weights corresponding to each implemented automotive project according to the time nodes of the implemented automotive projects is expressed as: ; where is the time increase and decrease weight corresponding to the k-th implemented automotive project, is the time node of the automotive innovation project, is the time node of the k-th implemented automotive project, is the technology attenuation cycle coefficient.

[0011] Optionally, obtaining the second correlation degree corresponding to every two reference words according to the time increase and decrease weights corresponding to each implemented automotive project and the multiple correlation times corresponding to every two reference words is expressed as: , ; where is the second correlation degree corresponding to the i-th reference word and the j-th reference word, is the number of implemented automotive projects, is the number of times the i-th reference word and the j-th reference word appear in the k-th implemented automotive project, is the time increase and decrease weight corresponding to the k-th implemented automotive project.

[0012] Optionally, obtaining the target credibility according to the first reference credibility and the second reference credibility is expressed as: , ; where is the target credibility, is the first reference credibility, is the second reference credibility, is the basic weight, is the attenuation weight.

[0013] There is also provided a knowledge graph-based credibility evaluation system for automotive innovation projects. The system is used to implement a knowledge graph-based credibility evaluation method for automotive innovation projects. The system includes: an acquisition module, which is used to obtain implemented automotive projects based on the knowledge graph, extract multiple automotive domain keywords from the implemented automotive projects to form an automotive keyword library, and obtain the text information of the automotive innovation project, obtain the automotive domain keywords appearing in the text information according to the text information and the automotive keyword library as benchmark words, and obtain the number of times each benchmark word appears in the text information as the independent frequency; a first evaluation module, which is used to obtain the first correlation degree of every two benchmark words in the implemented automotive projects, and obtain the first benchmark credibility corresponding to the automotive innovation project based on the benchmark model, the first correlation degree, and multiple independent frequencies; a second evaluation module, which is used to obtain the time nodes of each implemented automotive project, and obtain the second correlation degree of every two benchmark words in the implemented automotive projects based on the time nodes of each implemented automotive project, and obtain the second benchmark credibility corresponding to the automotive innovation project based on the benchmark model, the second correlation degree, and multiple independent frequencies; an evaluation decision module, which is used to obtain the target credibility according to the first benchmark credibility and the second benchmark credibility, and associate the target credibility with the automotive innovation project.

[0014] Optionally, the first evaluation module is further used to: obtain the number of times every two benchmark words appear simultaneously in each implemented automotive project and record it as the correlation number; obtain the first correlation degree corresponding to every two benchmark words according to the multiple correlation numbers corresponding to every two benchmark words.

[0015] The beneficial effects of the present invention are reflected in: In the entire knowledge graph-based credibility evaluation method for automotive innovation projects, through the deep integration of the knowledge graph and the multi-dimensional dynamic evaluation model, the core problems in the traditional credibility evaluation of automotive innovation projects, such as the staticization of technical correlation, the lack of time evolution characteristics, and the interference of high-frequency words, are effectively solved, and the accuracy, timeliness, and adaptability of the evaluation results are significantly improved. Among them, a dynamically updated automotive domain keyword library is first constructed. On this basis, a dual-path correlation degree calculation model is adopted: the first benchmark credibility focuses on the stability analysis of historical technology combinations, and the second benchmark credibility introduces a time increase and decrease weight mechanism, assigns higher weights to recent technology hotspots, and at the same time attenuates and suppresses obsolete technology combinations. The two generate the target credibility through a configurable fusion strategy, which not only retains the stability evaluation of mature technology solutions but also strengthens the forward-looking prediction of hot innovations. Description of the Drawings

[0016] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following will briefly introduce the drawings required for the description of the specific embodiments or the prior art. In all the drawings, similar elements or parts are generally identified by similar reference numerals. In the drawings, the elements or parts are not necessarily drawn to scale.

[0017] Figure 1 Schematic diagram of the steps of the method for evaluating the credibility of automotive innovation projects based on a knowledge graph according to the present invention; Figure 2 Partial schematic diagram of step S2 in the method for evaluating the credibility of automotive innovation projects based on a knowledge graph according to the present invention; Figure 3 Partial schematic diagram of step S3 in the method for evaluating the credibility of automotive innovation projects based on a knowledge graph according to the present invention. Specific embodiments

[0018] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. Usually, the components of the embodiments of the present invention described and shown in the accompanying drawings here can be arranged and designed in various different configurations.

[0019] Therefore, the following detailed description of the embodiments of the present invention provided in the drawings is not intended to limit the scope of the claimed present invention, but merely represents selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0020] It should be noted that: similar reference numerals and letters represent similar items in the following drawings. Therefore, once an item is defined in one drawing, it does not need to be further defined and explained in subsequent drawings. In addition, the terms "first", "second", etc. are only used for distinguishing descriptions and cannot be understood as indicating or implying relative importance.

[0021] As Figure 1 shown, a method for evaluating the credibility of automotive innovation projects based on a knowledge graph is provided, including: S1. Obtain the implemented automotive projects based on the knowledge graph, extract multiple automotive domain keywords from the implemented automotive projects to form an automotive keyword library, and obtain the text information of the automotive innovation projects. Obtain the automotive domain keywords that appear in the text information according to the text information and the automotive keyword library as the reference words, and obtain the number of times each reference word appears in the text information as the independent frequency; S2. Obtain the first correlation degree of every two benchmark words in the implemented automotive projects, and obtain the first benchmark credibility corresponding to the automotive innovation project based on the benchmark model, the first correlation degree, and multiple independent frequencies; S3. Obtain the time nodes of each implemented automotive project, and obtain the second correlation degree of every two benchmark words in the implemented automotive projects based on the time nodes of each implemented automotive project, and obtain the second benchmark credibility corresponding to the automotive innovation project based on the benchmark model, the second correlation degree, and multiple independent frequencies; S4. Obtain the target credibility according to the first benchmark credibility and the second benchmark credibility, and associate the target credibility with the automotive innovation project.

[0022] In this embodiment, it should be noted that in S1, based on keyword extraction and frequency statistics of the knowledge graph, a dynamic keyword library in the automotive field is constructed and the technical features of the innovation project are quantified. Specifically, first, all the data of the implemented automotive projects (such as technical documents, patents, or R & D reports) are retrieved from the knowledge graph, and the core terms in the project description are extracted through natural language processing technology, such as "in-vehicle sensor", "hydrogen fuel cell", "by-wire chassis", etc., to form an automotive keyword library covering technical nodes, components, and processes. This library ensures strong relevance of the keywords to the automotive innovation scenario through semantic disambiguation and domain filtering. For example, for new energy projects, the library may contain subdivision technical terms such as "solid-state battery thermal management" and "charging pile compatibility", while traditional structure projects cover terms such as "body stiffness topology optimization" and "powertrain NVH", thus constructing a multi-dimensional technical semantic network.

[0023] Furthermore, for the automotive innovation project to be evaluated, parse its text information (such as technical solution, requirement specification), match all associated benchmark words based on the keyword library, and count the original occurrence frequency of each benchmark word in the text. For example, in the text of a certain intelligent driving project, the benchmark word "multi-modal fusion algorithm" appears with a relatively high frequency, indicating that this technology is the core of the project; while the frequency of "lidar calibration" is relatively low, which may belong to the auxiliary module. This process not only captures the distribution density of technical elements but also provides basic parameters for subsequent correlation degree calculation. It should be noted that the dynamic update mechanism of the keyword library can incorporate emerging technical terms (such as "vehicle-road-cloud integration") to avoid the problem of semantic library lag caused by technological iteration and ensure that the benchmark words can accurately reflect the technological hotspots of the innovation project.

[0024] In S2, based on the credibility calculation of static correlation degree and frequency suppression, through historical technical relevance analysis and frequency optimization processing, the technical rationality of automotive innovation projects is quantified. Specifically, first, for all pairwise combinations of benchmark words, their co-occurrence characteristics in implemented automotive projects are extracted. For example, the co-occurrence pattern of "autopilot algorithm" and "high-precision map" in different project documents. By aggregating the co-occurrence statistical values in historical data, the first correlation degree representing the stability of the technical combination is generated. This correlation degree reflects the technical synergy intensity between benchmark words. For example, if a technical combination co-occurs frequently in most historical projects, it indicates that its relevance is universal. Conversely, if it only appears sporadically, it may belong to a marginal combination. Further, in the benchmark model, the independent frequency is denoised through a suppression function (such as logarithmic compression or threshold truncation) to avoid the interference of high-frequency but low-correlation words (such as the general term "in-vehicle system") on the credibility assessment, thereby extracting the frequency contribution with actual technical value.

[0025] For example, when evaluating an electrification platform project, the benchmark words "thermal runaway warning of battery module" and "BMS redundant design" may show a high co-occurrence correlation degree in historical projects, indicating a strong technical coupling between them; while the "fast charging protocol compatibility" has a high independent frequency, but its weight is reasonably reduced after suppression to avoid exaggerating the credibility due to a single high-frequency word. By multiplying the weighted sum of the correlation degree mean and the suppressed frequency, this model not only strengthens the synergy effect of the technical combination but also weakens the noise impact of redundant words. The finally output first benchmark credibility can effectively reflect the advantages of the project in terms of technical maturity and logical coherence.

[0026] In S3, based on the correlation degree correction and credibility optimization of dynamic time weights, a time evolution mechanism is introduced to dynamically calibrate the technical relevance to improve the adaptability of the evaluation results to technical trends. Specifically, by analyzing the time distribution characteristics of implemented automotive projects, different weight coefficients are assigned to projects in different historical stages. For example, the time weight of projects in the past three years is relatively high, while the weight of projects five years ago is significantly reduced. On this basis, the co-occurrence relationship between benchmark words is calculated with time weights. For example, the high-frequency co-occurrence of "vehicle networking security protocol" and "OTA upgrade" in recent projects will obtain a higher weight, while the combination of "mechanical power steering" and "traditional gearbox" in early projects will have its weight attenuated due to technological obsolescence, thus dynamically correcting the timeliness deviation of the correlation degree. This mechanism enables the correlation intensity of the technical combination to reflect the true state of current industry technology iteration and avoid evaluation distortion caused by the lag of historical data.

[0027] Furthermore, when calculating the second benchmark credibility, by combining the time-weighted mean of the correlation degree with the independent frequency after suppression processing, an evaluation index that takes into account both technical synergy and timeliness is formed. For example, in a smart cockpit project, although the "multi-screen interaction design" and "AR-HUD" had a low correlation degree in the early projects, recent technological breakthroughs have made them frequently co-occur in the latest projects. The introduction of time weights has significantly increased the contribution value of their correlation degree; while for obsolete technologies such as the "CD player module", even if the independent frequency is high, its weight will be greatly weakened by the suppression function. This dynamic adjustment enables the evaluation model to automatically capture the law of technological hotspot migration, such as the transition trend from "ternary lithium batteries" to "solid-state batteries" in the new energy field, so as to more accurately judge the technological foresight and implementation feasibility of innovative projects.

[0028] In S4, for multi-dimensional credibility fusion and evaluation result generation, the core lies in integrating static technical correlation and dynamic time evolution characteristics to form an objective credibility that comprehensively reflects the technological feasibility and trend adaptability of innovative projects. Specifically, by performing weighted fusion on the first benchmark credibility (based on historical co-occurrence correlation and frequency suppression) and the second benchmark credibility (based on time-weighted correlation and frequency optimization), for example, by assigning a higher weight to the second benchmark credibility to highlight the impact of technological trends, so as to balance the evaluation deviation between technological maturity and foresight. For example, in a hybrid power project, if the "energy recovery system" and "electronic control clutch module" in its technical solution show medium performance in the first benchmark credibility, but their combination has a significantly increased correlation degree in recent projects and the second benchmark credibility is relatively high, the target credibility after fusion will be closer to the current mainstream technological direction, avoiding the underestimation risk caused by outdated historical data.

[0029] Furthermore, the generation mechanism of the target credibility can dynamically adjust the fusion strategy in combination with domain requirements. For example, for disruptive innovation projects (such as the research and development of all-solid-state batteries), if the correlation degree of their technical keywords has increased rapidly in recent projects, even if the historical co-occurrence frequency is low, by increasing the weight coefficient of the second benchmark credibility, the target credibility can still accurately reflect its technological breakthrough potential; while for projects that improve mature technologies (such as optimizing the energy efficiency of traditional engines), the stability evaluation of the first benchmark credibility is emphasized. Finally, the target credibility is bound to the project in the form of a numerical value or a grade. For example, mapping high-credibility results to "clear technical path and low implementation risk", providing a quantitative basis for investment decisions or the division of R & D priorities. At the same time, after quantification, it can support cross-project horizontal comparison, such as identifying the technical credibility differences between two types of projects: "vehicle-road collaboration" and "automatic driving algorithm", driving the precise allocation of resources.

[0030] In summary, in the entire credibility evaluation method for automotive innovation projects based on the knowledge graph, through the deep integration of the knowledge graph and the multi-dimensional dynamic evaluation model, the core problems in the credibility evaluation of traditional automotive innovation projects, such as the static technical correlation, the lack of time evolution characteristics, and the interference of high-frequency words, are effectively solved, and the accuracy, timeliness, and adaptability of the evaluation results are significantly improved. Among them, a keyword library for the automotive field that is dynamically updated is first constructed. On this basis, a dual-path correlation calculation model is adopted: the first benchmark credibility focuses on the stability analysis of historical technical combinations, and the second benchmark credibility introduces a time increase and decrease weight mechanism, assigning higher weights to recent technical hotspots and suppressing the attenuation of obsolete technical combinations at the same time. The two generate the target credibility through a configurable fusion strategy, which not only retains the stability evaluation of mature technical solutions but also strengthens the forward-looking prediction of hot innovations.

[0031] As Figure 2 shown, in one embodiment, obtaining the first correlation degree of every two benchmark words in the implemented automotive projects in S2 includes: S21. Obtain the co-occurrence times of every two benchmark words in each implemented automotive project and record them as the correlation times; S22. Obtain the first correlation degree corresponding to every two benchmark words according to the multiple correlation times corresponding to every two benchmark words.

[0032] In this embodiment, it should be noted that in S21, specifically, for each pair of benchmark words, all implemented automotive project documents (such as patents, technical reports) are traversed, and the co-occurrence times of the two in the same project are counted. For example, the co-occurrence frequency of "autopilot algorithm" and "multi-sensor fusion" in a certain intelligent driving project, or the co-occurrence situation of "battery thermal management" and "BMS control" in a new energy project. This process is automatically completed through text mining technology. For example, the context co-occurrence relationship of technical terms is identified using a sliding window or semantic dependency analysis to avoid misjudgment caused by simple position proximity. For example, in the analysis of a certain vehicle-mounted communication project, "V2X protocol" and "low-latency transmission" may co-occur in multiple paragraphs due to the functional coupling in the technical solution description, while the co-occurrence times of "in-vehicle entertainment system" and "battery module" are zero because of their technical irrelevance, thus accurately capturing the true relevance of the technical combination.

[0033] In S22, it is the stage of correlation integration and normalization, and its goal is to transform the scattered co-occurrence features into standardized technical correlation indicators. Specifically, aggregate analysis is performed on the co-occurrence times of each pair of benchmark words in all historical projects. For example, calculate their average co-occurrence intensity in different projects. If a certain technology combination (such as "steer-by-wire" and "redundant control system") stably co-occurs in most projects, the average value is relatively high; conversely, if it only appears occasionally in a single project, the average value is relatively low. Further through normalization processing, the numerical value is mapped to the interval from 0 to 1. For example, when the total co-occurrence times of a pair of benchmark words in historical projects occupy the highest level in the entire library, its normalized result approaches 1, indicating a very strong technical correlation; while approaching 0 reflects the marginalization of the combination. This processing makes the correlation degrees in different technical fields comparable. For example, the correlation degree between "hydrogen fuel cell stack" and "proton exchange membrane" in the new energy field can be horizontally compared with the correlation degree between "vehicle body lightweighting" and "carbon fiber material" in the structural design field, providing a unified benchmark for the evaluation of cross-field innovation projects.

[0034] In one implementation, in S22, the first correlation degree corresponding to each two benchmark words is obtained according to the multiple correlation times corresponding to each two benchmark words, which is expressed as: , ; where is the first correlation degree corresponding to the i-th benchmark word and the j-th benchmark word, is the number of implemented automotive projects, is the number of occurrences of the i-th benchmark word and the j-th benchmark word in the k-th implemented automotive project.

[0035] In this implementation, it should be noted that in the expression , it represents the average co-occurrence times of two benchmark words in all implemented automotive projects. For example, if the benchmark word pair ("autopilot algorithm", "high-precision map") co-occurs 100 times in 100 historical projects, the average value is 1 time / project. In this way, the influence of the difference in the number of projects can be eliminated. Specifically, by dividing by the total number of projects , statistical deviation caused by the number of projects can be avoided. For example, if a certain technology combination co-occurs 10 times in 10 projects (average 1 time / project) and co-occurs 100 times in 100 projects (average 1 time / project), their correlations should be equivalent, rather than the latter being misjudged as a stronger correlation because of the higher total number; it can also quantify the universality of the technology combination. Specifically, the average value reflects the universality of the technology combination in historical projects. High-frequency co-occurrence (such as an average of 5 times / project) indicates stable technical collaboration, while low-frequency (such as an average of 0.1 times / project) may be an accidental combination.

[0036] Furthermore, Normalization is achieved, that is, the output range of the average co-occurrence times is compressed to (0, 1). In this way, when the average number is low (such as x = 1), the output value increases slowly (about 0.73); when the number is high (such as x = 5), the output approaches 1 but the growth rate slows down significantly (about 0.99), avoiding the over-amplification of high-frequency but low-value words (such as the general term "vehicle-mounted system") due to their high absolute frequency; at the same time, the correlation degrees of all technology combinations are mapped to the same scale (0, 1), facilitating horizontal comparison. For example, "battery thermal management - BMS control" in the new energy field and "body stiffness - lightweight materials" in the traditional field can be evaluated by the same standard.

[0037] To sum up, if the total frequency or linear normalization is directly used, high-frequency but weakly technologically related words (such as "vehicle-mounted system" frequently appearing in multiple projects) will lead to a falsely high correlation degree. Through this expression, even if a pair of benchmark words co-occur extremely frequently in a few projects (such as = 50 in a second project), their contribution will be diluted by the average value (for example, when n = 100, the contribution of a single project only increases by 0.5), and the increase in the correlation degree is limited after normalization and compression.

[0038] Suppose two pairs of benchmark words: Benchmark word pair A: The average co-occurrence is 10 times in 10 projects (total frequency 100), then = 10, ≈ 1.0. Benchmark word pair B: The average co-occurrence is 1 time in 100 projects (total frequency 100), then = 1, ≈ 0.73. Although the total frequencies are the same, benchmark word pair A is determined to be strongly correlated because of the high frequency concentrated in a few projects, while benchmark word pair B is reasonably weakened because of the widespread low-frequency co-occurrence.

[0039] In one embodiment, in S2, the benchmark model in obtaining the first benchmark credibility corresponding to the automotive innovation project based on the benchmark model, the first correlation degree, and multiple independent frequencies is expressed as: ; where is the first benchmark credibility, is the first correlation degree corresponding to the i-th benchmark word and the j-th benchmark word, is the number of the first correlation degrees, is the independent frequency corresponding to the P-th benchmark word, is the number of benchmark words.

[0040] In this embodiment, it should be noted that represents the average value of the first correlation degrees of all benchmark word pairs. For example, if a project contains 3 benchmark words (A, B, C), then it is necessary to calculate , , There are a total of 3 pairs of correlation degrees, and their average value is calculated. In this way, first, the synergy balance level of all technical combinations in the project can be reflected through the average value. For example, if the average correlation degree of a project is 0.9 (close to 1), it indicates that most combinations in its technical solution have strong historical synergy; if the average value is only 0.3, the technical logic may be loose or rely on marginal combinations. Second, the deviation in the number of reference words can be eliminated. Dividing by the logarithm (the total number of the first correlation degree) can avoid the credibility deviation caused by the number of reference words in the project. For example, a project with 5 reference words needs to calculate 10 pairs of correlation degrees, while a project with 3 reference words only calculates 3 pairs. After averaging, the comparability is enhanced.

[0041] Furthermore, in the second part of the expression , non-linear compression of the independent frequency of the reference words is reflected. The numerator part is the original frequency of a single reference word ; the denominator part , where is the total frequency of all reference words; suppressing high , the logarithmic function grows slowly, so that the higher the total frequency, the greater the increase in the denominator, thereby compressing the contribution of each ; at the same time, retaining low effective signals. Regardless of how the total frequency changes, a certain weight can still be retained to avoid the effective signals in the technical sub-field being completely submerged.

[0042] Furthermore, finally, the average correlation degree is multiplied by the sum of the suppressed frequencies, that is = average correlation degree × sum of suppressed frequencies. First, the average correlation degree reflects the synergy of the technical combination, and the sum of the suppressed frequencies reflects the density of technical elements. Multiplying the two ensures that the credibility satisfies both "reasonable technical logic" and "prominent core elements"; for example, if a project has a high average correlation degree but an extremely low sum of frequencies (the technical solution description is brief), or an extremely high sum of frequencies but a low average correlation degree (stacking high-frequency irrelevant terms), its credibility will be suppressed; the product operation strengthens the synergy effect of high correlation degree and reasonable frequencies. For example, the product of an average correlation degree of 0.8 and a sum of suppressed frequencies of 20 (16) is significantly higher than the product of a correlation degree of 0.5 and a sum of suppressed frequencies of 30 (15), reflecting that technical synergy takes precedence over simple element stacking.

[0043] For example: Project A: The total frequency is 1000, and the sum of the suppressed frequencies = 1000 / 7.9 ≈ 126.6. Project B: The total frequency is 200, and the sum of the suppressed frequencies = 200 / 6.3 ≈ 31.7. Although the total frequency of Project A is 5 times higher, the suppressed frequency is only 4 times different, and further constraints are combined with the average correlation degree to avoid a single high-frequency word dominating the evaluation.

[0044] Further explanation: To evaluate the credibility of an intelligent driving project, it includes 3 benchmark words (A: multi-sensor fusion, B: environmental perception algorithm, C: redundant control), with independent frequencies F1 = 50, F2 = 30, F3 = 20, and correlation degrees = 0.9, = 0.7, = 0.8. Then the average correlation degree is 0.8, and the total frequency is 100, 17.8, and finally the credibility of the first benchmark is obtained as 14.24.

[0045] Comparison scenario: If another project includes 3 benchmark words (A: battery thermal management, B: autonomous driving algorithm, C: in-vehicle entertainment), with inflated independent frequencies (F1 = 200, F2 = 150, F3 = 150, total frequency = 500), but a low average correlation degree ( = 0.1, = 0.1, = 0.1, average = 0.1). Then 69.25, and finally the credibility of the first benchmark is obtained as 6.9.

[0046] Then for Project 1: Although the frequency is low, it obtains a reasonable credibility (14.24) due to the high correlation degree.

[0047] Project 2: Despite the inflated frequency, due to the low average correlation degree, the credibility (6.9) is not higher than that of Project 1. Comparison conclusion: The technical solution of Project 1 (high synergy + reasonable frequency) is superior to that of Project 2 (low synergy + high-frequency stacking).

[0048] As Figure 3 shown, in one implementation, obtaining the second correlation degree of each two benchmark words in the implemented automotive projects based on the time nodes of each implemented automotive project in S3 includes: S31. Obtain the number of times each two benchmark words appear simultaneously in each implemented automotive project and record it as the correlation times; S32. Obtain the time increase and decrease weights corresponding to each implemented automotive project according to the time nodes of each implemented automotive project; S33. Obtain the second correlation degree corresponding to each two benchmark words according to the time increase and decrease weights corresponding to each implemented automotive project and the multiple correlation times corresponding to each two benchmark words.

[0049] In this embodiment, it should be noted that in S31, for each pair of reference words, all implemented automotive project documents are traversed to count the co-occurrence times of the two in the same project. For example, the co-occurrence frequency of "vehicle networking protocol" and "OTA upgrade" in a certain intelligent network connection project. Similar to the static statistics in S21, the difference is that the association times in S31 will be used as the second association degree calculated in combination with time increase and decrease weights in S33. For example, a pair of reference words may co-occur frequently in early projects but have a reduced frequency in recent projects. In this case, the original data needs to be retained to support dynamic correction. This step ensures the integrity of historical data through co-occurrence statistics independent of time sequence, avoiding information loss caused by pre-weighting.

[0050] In S32, the association times are weighted differentially according to the proximity of the project time. Specifically, a time increase and decrease weight is assigned to each implemented automotive project. For example, the weight of a project within the last 3 years is 1.8, and the weight of a project 5 years ago is 0.5. The weight values show a non-linear distribution as they decay over time (such as exponential decay). The weight range is limited between 0 and 2, which not only retains the reference value of historical data (the weight is not 0) but also highlights the priority of recent projects. For example, for the technology combination of "solid-state battery" and "thermal runaway warning", if it only appears frequently in projects in the last two years, the high weight of recent projects will significantly increase its association degree; while for the combination of "mechanical supercharger" and "carburetor" in early projects, due to technological obsolescence, its weight approaches 0, thus suppressing the interference of obsolete technologies on the evaluation.

[0051] In S33, the time weight and the co-occurrence times are integrated to generate a dynamically corrected second association degree. For example, the association times of "AR-HUD" and "multi-screen interaction" were low and the weight decayed in early projects, while in recent projects, due to technological breakthroughs, the association times increased sharply and the weight was high. After integration, its second association degree is significantly higher than the static association degree. This mechanism enables the association strength of technology combinations to dynamically reflect industry trends. For example, in the field of electrification, the transition from "ternary lithium battery" to "cobalt-free battery", even if the historical co-occurrence times are low, the increase in recent weights can still drive the association degree to rise, thus accurately identifying the collaborative value of emerging technologies.

[0052] In one embodiment, the time increase and decrease weights corresponding to each implemented automotive project are obtained according to the time nodes of each implemented automotive project in S32, which is expressed as: ; where is the time increase and decrease weight corresponding to the k-th implemented automotive project, is the time node of the automotive innovation project, is the time node of the k-th implemented automotive project, is the technology decay period coefficient.

[0053] In this embodiment, it should be noted that Among them represents the time node of the current automotive innovation project and the time node of the k-th historical implemented automotive project The time difference (in years); as the time difference increases, the exponential term The value of linearly increases, causing the denominator to expand rapidly, thereby making the weight Exponentially decay. In this way, the weight of recent projects (small time difference) is significantly higher than that of long-term projects. For example, if - = 0 (i.e., the historical project and the innovation project are in the same year), the weight = 2 (the maximum value); if the time difference is 10 years ( = 0.1), the weight decays to 0.736. More importantly, The decay rate can be controlled. The cycle of general technology is 7 years, then Can be 0.1, indicating that the half-life of the technical correlation is about 7 years (the time difference when the weight drops to 1 is 7 years), which is in line with the characteristics of the relatively long iteration cycle of automotive technology.

[0054] In summary, the obsolete technology combinations in early projects (such as "carburetor - supercharger") have a falsely high correlation in the static model due to their high historical co-occurrence frequency. This embodiment can suppress obsolete technologies through time decay; it can also solve the problem that emerging technology combinations (such as "solid-state battery - thermal management") do not exist or appear with low frequency in early projects, resulting in an underestimated static correlation.

[0055] In one embodiment, in S33, the second correlation degree corresponding to each two reference words is obtained according to the time increase and decrease weight corresponding to each implemented automotive project and the multiple association times corresponding to each two reference words, which is expressed as: , ; where Is the second correlation degree corresponding to the i-th reference word and the j-th reference word, Is the number of implemented automotive projects, Is the number of occurrences of the i-th reference word and the j-th reference word in the k-th implemented automotive project, Is the time increase and decrease weight corresponding to the k-th implemented automotive project.

[0056] In this embodiment, it should be noted that, which is roughly the same as the expression for obtaining the first correlation degree, the difference is that in In [description], it represents the time-weighted average of the co-occurrence times of each pair of reference words in all historical projects. For example, if the reference word pair ("solid-state battery", "thermal management") co-occurs 10 times in the 2023 project (weight 2.0) and 5 times in the 2020 project (weight 1.48), the weighted average is 13.7. In this way, the recent technological trends can be strengthened: through the time weights , the co-occurrence times of recent projects are amplified. For example, the weight of the 2023 project is 1.35 times that of the 2020 project (2.0 / 1.48), and its contribution ratio of co-occurrence times is higher; it can also suppress the interference of historical data. The weight of early projects (such as in 2010) may be as low as below 0.5. Even if their co-occurrence times are high, the contribution after weighting is significantly weakened. For example, if there are 20 co-occurrences in the 2010 project (weight 0.5), its effective contribution is only 10, while if there are 20 co-occurrences in the 2023 project (weight 2.0), the contribution is 40.

[0057] In one implementation, obtaining the target credibility according to the first reference credibility and the second reference credibility in S4 is expressed as: , ; where, is the target credibility, is the first reference credibility, is the second reference credibility, is the basic weight, is the decay weight.

[0058] In this implementation, it should be noted that the two-dimensional weight allocation mechanism The expression dynamically balances the technology maturity and trend adaptability through the weight coefficients (basic weight) and (decay weight). The basic weight : reflects the importance of the static correlation degree ( ), emphasizing the stability of historical technology collaboration; for example, for an improvement project of a mature technology but with low recent popularity (such as the optimization of traditional engine energy efficiency), set = 0.8, emphasizing the reliability of historical data. The decay weight reflects the importance of the dynamic correlation degree ( ), emphasizing the forward-looking nature of technology trends; for example, for an innovative project with immature technology and low recent popularity (such as the research and development of all-solid-state batteries), set = 0.9. Even if its historical correlation degree is low, the credibility can still be improved through recent trends.

[0059] Furthermore, the weight coefficients clearly divide the contribution ratios of the two types of credibility, facilitating flexible adjustment according to evaluation requirements, and supporting the introduction of more evaluation dimensions (such as market risk, patent barriers) to maintain the consistency of the model framework. At the same time, the weight coefficients can be dynamically configured according to the technology life cycle (introduction period, growth period, maturity period). For example, in the introduction period = 0.9: Strengthen the emerging technology trend. In the maturity period = 0.8: Rely on the stability of historical technologies.

[0060] There is also provided a knowledge graph-based credibility evaluation system for automotive innovation projects. The system is used to implement the above-mentioned knowledge graph-based credibility evaluation method for automotive innovation projects. The system includes: An acquisition module, which is used to obtain implemented automotive projects based on the knowledge graph, extract multiple automotive domain keywords from the implemented automotive projects to form an automotive keyword library, and obtain the text information of the automotive innovation project. According to the text information and the automotive keyword library, the automotive domain keywords appearing in the text information are obtained as reference words, and the number of times each reference word appears in the text information is obtained as the independent frequency; A first evaluation module, which is used to obtain the first correlation degree of every two reference words in the implemented automotive projects, and obtain the first reference credibility corresponding to the automotive innovation project based on the reference model, the first correlation degree, and multiple independent frequencies; A second evaluation module, which is used to obtain the time nodes of each implemented automotive project, and obtain the second correlation degree of every two reference words in the implemented automotive projects based on the time nodes of each implemented automotive project, and obtain the second reference credibility corresponding to the automotive innovation project based on the reference model, the second correlation degree, and multiple independent frequencies; An evaluation decision module, which is used to obtain the target credibility according to the first reference credibility and the second reference credibility, and associate the target credibility with the automotive innovation project.

[0061] In one implementation, the first evaluation module is further used to: obtain the number of times every two reference words appear simultaneously in each implemented automotive project and record it as the associated number; obtain the first correlation degree corresponding to every two reference words according to the multiple associated numbers corresponding to every two reference words.

[0062] In one implementation, the second evaluation module is further used to: obtain the number of times every two reference words appear simultaneously in each implemented automotive project and record it as the associated number; obtain the time increase and decrease weights corresponding to each implemented automotive project according to the time nodes of each implemented automotive project; obtain the second correlation degree corresponding to every two reference words according to the time increase and decrease weights corresponding to each implemented automotive project and the multiple associated numbers corresponding to every two reference words.

[0063] In this embodiment, it should be noted that regarding the above-mentioned automobile innovation project credibility evaluation system based on the knowledge graph, the specific manner of performing operations has been described in detail in the embodiment of the method for evaluating the credibility of automobile innovation projects based on the knowledge graph, and will not be elaborated here.

[0064] The preferred embodiments of the present disclosure have been described in detail above in conjunction with the accompanying drawings. However, the present disclosure is not limited to the specific details in the above embodiments. Within the scope of the technical concept of the present disclosure, various simple modifications can be made to the technical solutions of the present disclosure, and these simple modifications all belong to the protection scope of the present disclosure.

[0065] In addition, it should be noted that, in the case of no contradiction, the various specific technical features described in the above specific embodiments can be combined in any appropriate manner. To avoid unnecessary repetition, the present disclosure will not separately describe various possible combination methods.

[0066] In addition, any combination can be made between various different embodiments of the present disclosure, as long as it does not violate the idea of the present disclosure, and it should also be regarded as the content disclosed by the present disclosure.

[0067] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements on some or all of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present invention, and they should all be covered by the scope of the claims and the description of the present invention.

Claims

1. A credibility assessment method for automobile innovation projects based on knowledge graph, characterized in that: include: Based on the knowledge graph, the implemented automobile projects are obtained, multiple automobile field keywords are extracted from the implemented automobile projects and form an automobile keyword library, and the text information of the automobile innovation projects is obtained. According to the text information and the automobile keyword library, the automobile field keywords appearing in the text information are obtained and used as benchmark words, and the number of times each benchmark word appears in the text information is obtained and used as an independent frequency; Obtaining the first correlation between every two benchmark words in the implemented automobile projects, and obtaining the first benchmark credibility corresponding to the automobile innovation project based on the benchmark model, the first correlation and a plurality of independent frequencies; Obtaining the time nodes of each implemented automobile project, and obtaining the second correlation between each two benchmark words in the implemented automobile project based on the time nodes of each implemented automobile project, and obtaining the second benchmark credibility corresponding to the automobile innovation project based on the benchmark model, the second correlation and multiple independent frequencies; The target credibility is obtained according to the first benchmark credibility and the second benchmark credibility, and the target credibility is associated with the automobile innovation project.

2. The method for evaluating the credibility of automobile innovation projects based on knowledge graph according to claim 1 is characterized in that: The obtaining of the first correlation between each two reference words in the implemented automobile project comprises: Get the number of times each two benchmark words appear together in each implemented automotive project and record it as the number of associations; A first degree of association corresponding to each two reference words is obtained according to a plurality of association times corresponding to each two reference words.

3. The method for evaluating the credibility of automobile innovation projects based on knowledge graph according to claim 2 is characterized in that: The first correlation degree corresponding to each two benchmark words is obtained according to the multiple correlation times corresponding to each two benchmark words as follows: , ;in, is the first correlation between the i-th benchmark word and the j-th benchmark word, is the number of implemented automotive projects, is the number of occurrences of the i-th benchmark word and the j-th benchmark word in the k-th implemented automotive project.

4. The method for evaluating the credibility of automobile innovation projects based on knowledge graph according to claim 1 is characterized in that: The benchmark model in the first benchmark credibility corresponding to the automobile innovation project obtained based on the benchmark model, the first correlation and multiple independent frequencies is expressed as: ;in, is the first benchmark credibility, is the first correlation between the i-th benchmark word and the j-th benchmark word, is the number of first degree of association, is the independent frequency corresponding to the Pth benchmark word, is the number of base words.

5. The method for evaluating the credibility of automobile innovation projects based on knowledge graph according to claim 1 is characterized in that: The step of obtaining the second correlation between each two reference words in the implemented automobile project based on the time nodes of each implemented automobile project comprises: Get the number of times each two benchmark words appear together in each implemented automotive project and record it as the number of associations; According to the time nodes of each implemented automobile project, the time increase and decrease weights corresponding to each implemented automobile project are obtained; The second correlation degree corresponding to every two reference words is obtained according to the time increase / decrease weight corresponding to each implemented automobile project and a plurality of correlation times corresponding to every two reference words.

6. The method for evaluating the credibility of automobile innovation projects based on knowledge graph according to claim 5 is characterized in that: The time increase and decrease weights corresponding to each implemented automobile project are obtained according to the time nodes of each implemented automobile project as follows: ;in, is the time increase or decrease weight corresponding to the kth implemented automobile project, The time node for automotive innovation projects, is the time node of the kth implemented automotive project, is the technology decay period coefficient.

7. The method for evaluating the credibility of automobile innovation projects based on knowledge graph according to claim 5 is characterized in that: The second correlation degree corresponding to each two benchmark words is obtained according to the time increase and decrease weight corresponding to each implemented automobile project and the multiple correlation times corresponding to each two benchmark words as follows: , ;in, is the second correlation between the i-th benchmark word and the j-th benchmark word, is the number of implemented automotive projects, is the number of occurrences of the i-th benchmark word and the j-th benchmark word in the k-th implemented automotive project, Increase or decrease the weight of the time corresponding to the k-th implemented automobile project.

8. The method for evaluating the credibility of automobile innovation projects based on knowledge graph according to claim 1 is characterized in that: The target credibility is obtained according to the first reference credibility and the second reference credibility as follows: , ;in, is the target credibility, is the first benchmark credibility, is the second benchmark credibility, is the basic weight, is the attenuation weight.

9. A knowledge graph-based automobile innovation project credibility assessment system, characterized in that: The system is used to implement the automobile innovation project credibility assessment method based on knowledge graph as described in any one of claims 1 to 8, and the system includes: An acquisition module is used to acquire implemented automobile projects based on the knowledge graph, extract multiple automobile field keywords from the implemented automobile projects and form an automobile keyword library, and acquire text information of automobile innovation projects, acquire automobile field keywords appearing in the text information according to the text information and the automobile keyword library and use them as benchmark words, and acquire the number of times each benchmark word appears in the text information and use it as an independent frequency; A first evaluation module is used to obtain a first correlation between every two benchmark words in implemented automobile projects, and obtain a first benchmark credibility corresponding to the automobile innovation project based on the benchmark model, the first correlation and a plurality of independent frequencies; A second evaluation module is used to obtain the time nodes of each implemented automobile project, and based on the time nodes of each implemented automobile project, obtain the second correlation between every two benchmark words in the implemented automobile project, and obtain the second benchmark credibility corresponding to the automobile innovation project based on the benchmark model, the second correlation and multiple independent frequencies; The evaluation decision module is used to obtain the target credibility according to the first benchmark credibility and the second benchmark credibility, and associate the target credibility with the automobile innovation project.

10. The automobile innovation project credibility assessment system based on knowledge graph according to claim 9 is characterized in that: The first evaluation module is also used for: Get the number of times each two benchmark words appear together in each implemented automotive project and record it as the number of associations; A first degree of association corresponding to each two reference words is obtained according to a plurality of association times corresponding to each two reference words.

Images