A method, device, equipment and medium for evaluating enterprise technology achievements

Abstract

The application discloses an enterprise technology achievement evaluation method and device, equipment and medium, and relates to the technical field of computers. The method comprises the following steps: acquiring multi-source data related to a target enterprise technology achievement, wherein the multi-source data comprises technical dimension data and market dimension data; performing feature fusion operation on the technical dimension data and the market dimension data by using a preset space-time alignment engine and a preset semantic alignment engine to obtain a fusion feature vector; constructing a technology node and a market node based on the fusion feature vector, creating a cross-domain attention network based on a bidirectional attention mechanism, and constructing a dynamic correlation network graph based on the technology node and the market node; and analyzing the dynamic correlation network graph by using a preset deep learning algorithm to obtain a target analysis result corresponding to the target enterprise technology achievement. Thus, the problem of data fragmentation, dynamicity loss and insufficient cross-domain correlation can be solved, and the joint intelligent prediction of technology maturity and market adaptability can be realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of computer, in particular to a technology achievement evaluation method and device of enterprise, equipment and medium. BACKGROUND

[0002] Technology readiness level (TRL) is a standard for measuring and evaluating the maturity of technology. At present, there are the following problems in the evaluation of technology maturity and market adaptability of technology achievements in the process of transformation to the market: technology evaluation and market evaluation are usually independent of each other, and there is a lack of effective cross-domain data correlation and comprehensive analysis method. Traditional technology maturity evaluation methods often focus only on technical indicators in the research and development stage, and fail to fully incorporate market demand dynamics for feedback adjustment. Existing evaluation models are mostly static models, which are difficult to capture and respond to sudden changes in the technology evolution process and dramatic fluctuations in the market environment such as sudden adjustments in policies and regulations. Traditional regression models that rely on historical data have high prediction error rates when evaluating emerging technology fields, making it difficult to adapt to rapidly changing environments. Existing correlation technologies such as knowledge graphs are mainly applied in a single field, and lack a mechanism for establishing a deep and effective correlation between technology and the market.

[0003] From the above, how to solve the problems of data fragmentation, lack of dynamics and insufficient cross-domain correlation, and realize the joint intelligent prediction of technology maturity and market adaptability is an urgent problem to be solved. SUMMARY

[0004] Therefore, the purpose of the present application is to provide a technology achievement evaluation method and device of enterprise, equipment and medium, which can solve the problems of data fragmentation, lack of dynamics and insufficient cross-domain correlation, and realize the joint intelligent prediction of technology maturity and market adaptability. The specific solutions are as follows:

[0005] In a first aspect, the present application provides a technology achievement evaluation method of enterprise, comprising:

[0006] Obtaining multi-source data related to the target enterprise technology achievement, including technology dimension data and market dimension data;

[0007] Performing feature fusion operation on the technology dimension data and the market dimension data by using a preset spatio-temporal alignment engine and a preset semantic alignment engine to obtain a fusion feature vector; the technology dimension data includes patent data, paper data and research and development data; the market dimension data includes user behavior data, industry report data and competitive intelligence data; the fusion feature vector includes a technology feature vector and a market feature vector;

[0008] construct a technology node and a market node based on the fusion feature vector, and utilize a cross-domain attention network created based on a bidirectional attention mechanism, and construct a dynamic correlation network graph based on the technology node and the market node; the dynamic correlation network graph is a graph representing correlation between the technology node and the market node;

[0009] analyze the dynamic correlation network graph by using a preset deep learning algorithm to obtain a target analysis result corresponding to the technical achievement of the target enterprise; the target analysis result includes a technology maturity level, a market adaptation index, and a coordination risk probability.

[0010] Optionally, the feature fusion operation on the technology dimension data and the market dimension data by using the preset spatio-temporal alignment engine and the preset semantic alignment engine to obtain a fusion feature vector includes:

[0011] a mapping reference offset rule for a technology stage to a market window period is designed based on a dynamic offset mechanism by using the preset spatio-temporal alignment engine;

[0012] a dynamic offset amount related to a technology breakthrough strength is generated based on a preset compensation algorithm by using the preset spatio-temporal alignment engine, and the technology dimension data and the market dimension data are mapped onto a unified timeline based on the mapping reference offset rule and the dynamic offset amount to obtain the technology dimension data and the market dimension data after spatio-temporal alignment;

[0013] a feature fusion operation is performed on the technology dimension data and the market dimension data after spatio-temporal alignment by using the preset semantic alignment engine in a double-path vectorization manner to obtain a fusion feature vector.

[0014] Optionally, the construction of the technology node and the market node based on the fusion feature vector includes:

[0015] the technology feature vector is deepened based on an international patent classification tree to obtain a technology node;

[0016] the market feature vector is expanded based on an industry chain to obtain a market node.

[0017] Optionally, the utilization of the cross-domain attention network created based on the bidirectional attention mechanism and the construction of the dynamic correlation network graph based on the technology node and the market node include:

[0018] The cross-domain attention network is created based on a bidirectional attention mechanism, and a first association relationship representing the contribution of technology to the market and a second association relationship representing the traction of the market to technology are established based on the technology node and the market node, and a preset double attention mechanism is used to calculate a first association strength of the first association relationship and a second association strength of the second association relationship.

[0019] A dynamic association network graph is constructed based on the first association relationship, the second association relationship, the technology node and the market node.

[0020] Optionally, after the cross-domain attention network created based on the bidirectional attention mechanism and the dynamic association network graph constructed based on the technology node and the market node, the method further comprises:

[0021] Based on a preset dynamic maintenance strategy, technology nodes and market nodes with an association tightness lower than a preset strength threshold or a change amplitude of the association tightness greater than a preset change threshold within a preset time are removed to optimize the dynamic association network graph; the value of the association tightness is a geometric mean value of the first association strength and the second association strength.

[0022] Optionally, the method further comprises:

[0023] Market fluctuation monitoring data is obtained, and an evaluation weight parameter is adjusted based on the market fluctuation monitoring data;

[0024] The dynamic association network graph is analyzed based on the adjusted evaluation weight parameter using a preset deep learning algorithm to obtain a target analysis result corresponding to the target enterprise technology achievement.

[0025] Optionally, the enterprise technology achievement evaluation method further comprises:

[0026] The preset deep learning algorithm is updated based on a double-cycle dynamic updating mechanism; the double-cycle dynamic updating mechanism comprises a target short-term cycle update and a target long-term cycle update, the target short-term cycle update is a model update operation triggered based on a market fluctuation event, and the target long-term cycle update is a model update operation triggered based on a technology breakthrough event.

[0027] In a second aspect, the present application provides an enterprise technology achievement evaluation device, comprising:

[0028] A data acquisition module is configured to acquire multi-source data related to a target enterprise technology achievement, including technology dimension data and market dimension data.

[0029] The feature fusion module is used to perform feature fusion operations on the technical dimension data and the market dimension data using a preset spatiotemporal alignment engine and a preset semantic alignment engine to obtain a fused feature vector; the technical dimension data includes patent data, paper data, and R&D data; the market dimension data includes user behavior data, industry report data, and competitive intelligence data; the fused feature vector includes a technical feature vector and a market feature vector.

[0030] The graph construction module is used to construct technology nodes and market nodes based on the fused feature vectors, and to utilize a cross-domain attention network created based on a bidirectional attention mechanism to construct a dynamic correlation network graph based on the technology nodes and the market nodes; the dynamic correlation network graph is a graph representing the correlation between the technology nodes and the market nodes.

[0031] The results generation module is used to analyze the dynamic correlation network graph using a preset deep learning algorithm to obtain the target analysis results corresponding to the technological achievements of the target enterprise; the target analysis results include technology maturity level, market adaptability index and coordination risk probability.

[0032] Thirdly, this application provides an electronic device, comprising:

[0033] Memory, used to store computer programs;

[0034] A processor is used to execute the computer program to implement the aforementioned enterprise technology achievement evaluation method.

[0035] Fourthly, this application provides a computer-readable storage medium for storing a computer program, wherein the computer program, when executed by a processor, implements the aforementioned enterprise technology achievement evaluation method.

[0036] This application provides a method for evaluating enterprise technological achievements. First, it acquires multi-source data related to the target enterprise's technological achievements, including technology-dimensional data and market-dimensional data. Then, it uses a preset spatiotemporal alignment engine and a preset semantic alignment engine to perform feature fusion operations on the technology-dimensional data and the market-dimensional data to obtain a fused feature vector. The technology-dimensional data includes patent data, paper data, and R&D data; the market-dimensional data includes user behavior data, industry report data, and competitive intelligence data. The fused feature vector includes a technology feature vector and a market feature vector. Next, it constructs technology nodes and market nodes based on the fused feature vector, and uses a cross-domain attention network created based on a bidirectional attention mechanism to construct a dynamic correlation network graph based on the technology nodes and the market nodes. Finally, it uses a preset deep learning algorithm to analyze the dynamic correlation network graph to obtain the target analysis results corresponding to the target enterprise's technological achievements. The target analysis results include technology maturity level, market adaptability index, and coordination risk probability.

[0037] As can be seen from the above, this application performs feature fusion operations on the technical dimension data and the market dimension data through a preset spatiotemporal alignment engine and a preset semantic alignment engine, establishing a dynamic alignment mechanism between technical features and market features to eliminate data fragmentation; it utilizes a cross-domain attention network created based on a bidirectional attention mechanism, and constructs a dynamic correlation network graph based on the technical nodes and the market nodes to capture the implicit correlation between technological evolution and market demand; it uses a preset deep learning algorithm to analyze the dynamic correlation network graph to achieve joint intelligent prediction of technology maturity and market adaptability. Thus, it can solve the problems of data fragmentation, lack of dynamism, and insufficient cross-domain correlation, achieving joint intelligent prediction of technology maturity and market adaptability. Attached Figure Description

[0038] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0039] Figure 1 This is a flowchart of a method for evaluating enterprise technological achievements disclosed in this application;

[0040] Figure 2 This is a flowchart of a feature fusion architecture disclosed in this application;

[0041] Figure 3 This application discloses a flowchart of a graph construction architecture.

[0042] Figure 4 This is a flowchart of an intelligent evaluation architecture disclosed in this application;

[0043] Figure 5 This is a schematic diagram of an enterprise technology achievement evaluation device disclosed in this application;

[0044] Figure 6 This is a structural diagram of an electronic device disclosed in this application. Detailed Implementation

[0045] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0046] Technology maturity (TM) is a measure of the degree to which key technologies meet project objectives and is a crucial element of project risk. Technology maturity level (TMR) refers to a standard for measuring and evaluating technology maturity. Currently, the assessment of the technology maturity and market adaptability of scientific and technological achievements during market transformation faces the following problems: technology assessment and market assessment are usually conducted independently, lacking effective methods for cross-domain data correlation and comprehensive analysis. Traditional technology maturity assessment methods often focus only on technical indicators during the R&D stage, failing to fully incorporate dynamic market demand feedback for adjustment. Existing assessment models are mostly static, making it difficult to capture and respond in real time to sudden changes in technological evolution and drastic fluctuations in the market environment, such as sudden adjustments to policies and regulations. Traditional regression models relying on historical data have high prediction error rates when assessing emerging technology fields and are difficult to adapt to rapidly changing environments. Existing knowledge graph and other correlation technologies are mainly applied to single domains, lacking mechanisms for establishing deep and effective correlations between the technology and market domains. Therefore, this application provides a corporate technology achievement assessment scheme that can solve the problems of data fragmentation, lack of dynamism, and insufficient cross-domain correlation, achieving joint intelligent prediction of technology maturity and market adaptability.

[0047] See Figure 1 As shown in the embodiments of this application, a method for evaluating enterprise technological achievements is disclosed, including:

[0048] Step S11: Obtain multi-source data related to the target company's technological achievements, including technology-related data and market-related data.

[0049] In this embodiment, the technical dimension data includes, but is not limited to, patent data, paper data, and R&D data; the market dimension data includes, but is not limited to, user behavior data, industry report data, and competitive intelligence data. Multi-source data from technical dimensions such as patents, papers, and R&D data, and market dimensions such as user behavior, industry reports, and competitive intelligence are collected. The dynamic shift between technological stages and market windows is correlated through mapping, and the problem of data fragmentation is solved by semantically aligning patent terminology with market concepts.

[0050] Step S12: Use a preset spatiotemporal alignment engine and a preset semantic alignment engine to perform feature fusion operation on the technical dimension data and the market dimension data to obtain a fused feature vector.

[0051] See Figure 2As shown, this embodiment employs a spatiotemporal alignment and semantic alignment engine to bridge the spatiotemporal and semantic gap between technical data and market data. The fused feature vector includes a technical feature vector and a market feature vector. Specifically, the step of using a preset spatiotemporal alignment engine and a preset semantic alignment engine to perform feature fusion operations on the technical dimension data and the market dimension data to obtain a fused feature vector can include: using the preset spatiotemporal alignment engine to design a mapping benchmark offset rule based on a dynamic offset mechanism regarding the transition from a technical stage to a market window; using the preset spatiotemporal alignment engine to generate a dynamic offset related to the intensity of a technical breakthrough based on a preset compensation algorithm, and mapping the technical dimension data and the market dimension data to a unified timeline based on the mapping benchmark offset rule and the dynamic offset, to obtain spatiotemporally aligned technical dimension data and market dimension data; and using the preset semantic alignment engine to perform feature fusion operations on the spatiotemporally aligned technical dimension data and market dimension data using a dual-path vectorization approach to obtain a fused feature vector. In other words, a dynamic offset mechanism is used, a preset mapping benchmark offset rule for the transition from a technical stage to a market window is established, and a preset compensation algorithm is used to dynamically compress the offset based on the intensity of the technical breakthrough. Taking solid-state battery technology as an example, when the battery energy density exceeds 400Wh / kg, it signifies a major technological breakthrough, corresponding to a shortened market concept validation cycle. The above method maps technological events and market events onto a unified timeline. A dual-path vectorization approach is used to vectorize technical terms and market concepts. In terms of technical terms, IPC (International Patent Classification) numbers are embedded hierarchically for vectorization. For example, H01M10 / 0562 can be divided into H (part) Electrical, H01 (class) Basic Electrical Components, H01M (subclass) Batteries, H01M10 (group) Secondary Batteries, and H01M10 / 0562 (group) Solid-State Electrolyte Lithium Batteries. Furthermore, H01M10 / 0562 is encoded as a 128-dimensional vector. Regarding market concepts, an industry corpus can be built first, followed by fine-tuning of the BERT (Bidirectional Encoder Representations from Transformers, an open-source machine learning framework designed for natural language processing) model. The industry corpus can then be vectorized using BERT, encoding market demands into 768-dimensional vectors. This establishes a dynamic alignment mechanism between technical and market characteristics, eliminating data fragmentation.

[0052] Step S13: Construct technology nodes and market nodes based on the fused feature vectors, and utilize a cross-domain attention network created based on a bidirectional attention mechanism to construct a dynamic relational network graph based on the technology nodes and the market nodes.

[0053] In this embodiment, based on feature fusion, technology nodes and market nodes are constructed through a cross-domain attention network. Specifically, the construction of technology nodes and market nodes based on the fused feature vectors can include: deepening the technology feature vectors based on the International Patent Classification (IPC) tree to obtain technology nodes; and expanding the market feature vectors based on the industry chain to obtain market nodes. That is, technology nodes can be vertically split along the IPC classification tree to the fourth-level group number, gradually deepening. For example, the node deepening of solid-state batteries: H01M (battery) — H01M10 (rechargeable battery) — H01M10 / 0562 (solid electrolyte). Market nodes can be horizontally extended along the industry chain to the third-level entity. For example, lithium mine supplier — battery material manufacturer — new energy vehicle manufacturer.

[0054] Furthermore, establishing correlations, constructing a dynamically evolving technology-market correlation network, and capturing implicit relationships. Specifically, the method of utilizing a cross-domain attention network created based on a bidirectional attention mechanism, and constructing a dynamic correlation network graph based on the technology nodes and the market nodes, can include: using a cross-domain attention network created based on a bidirectional attention mechanism, and establishing a first correlation relationship representing the technology's contribution to the market and a second correlation relationship representing the market's pull on the technology based on the technology nodes and the market nodes; and using a preset dual attention mechanism to calculate the first correlation strength of the first correlation relationship and the second correlation strength of the second correlation relationship respectively; and constructing a dynamic correlation network graph based on the first correlation relationship, the second correlation relationship, the technology nodes, and the market nodes. That is, by replacing the traditional static correlation with a bidirectional attention mechanism, the correlation strength of the two parts—the technology's contribution to the market and the market's pull on the technology—is calculated, and the interaction between the technology and the market is quantified in real time. The technology's contribution to the market, i.e., the driving force of technological breakthroughs on market hotspots, can be calculated through patent growth rate and market attention, as shown in the following formula:

[0055] ;

[0056] in, The strength of the correlation between technology's contribution to the market. Market pull on technology, which assesses the strength of how changes in demand influence technological pathways, can be calculated using the rate of change in demand and the rate of change in technology maturity. The specific formula is as follows:

[0057] ;

[0058] in, The strength of the market's correlation with the driving force of technology; This is the rate of change in demand, used to represent the speed of market expansion. The technology maturity rate is used to represent the speed of technological evolution. This represents the number of related pairs with market-driven technological influence within the current calculation period.

[0059] Furthermore, this embodiment filters out weakly related nodes through dynamic maintenance and uses a weight decay method to update weights for intelligent pruning. Specifically, after constructing a dynamic relationship network graph based on the technology nodes and market nodes using a cross-domain attention network created based on a bidirectional attention mechanism, it may further include: removing technology nodes and market nodes with a relationship strength lower than a preset strength threshold or whose relationship strength changes more than a preset change threshold within a preset time period based on a preset dynamic maintenance strategy, thereby optimizing the dynamic relationship network graph; the relationship strength is the geometric average of the first relationship strength and the second relationship strength. That is, weakly related nodes with low coupling strength or excessively high monthly activity decreases are removed. This further reduces computational complexity and improves the accuracy of key relationship identification.

[0060] Step S14: Analyze the dynamic correlation network graph using a preset deep learning algorithm to obtain the target analysis results corresponding to the technological achievements of the target enterprise.

[0061] In this embodiment, the features of the constructed knowledge graph are dynamically analyzed using a deep learning algorithm to obtain the target analysis results. The target analysis results include technology maturity level, market fit index, and coordination risk probability. Figure 4 As shown, in one specific implementation, a multi-task Transformer model extracts cross-domain features through a shared encoder and outputs three intelligent evaluation indicators through dynamic weight adjustment. Simultaneously, considering the rapid market changes and complex influencing factors, real-time market node update data is integrated, and evaluation weight parameters are dynamically adjusted to jointly output an intelligent judgment of technological value assessment and market prospect prediction. Specifically, the step of using a preset deep learning algorithm to analyze the dynamic correlation network graph to obtain the target analysis result corresponding to the target enterprise's technological achievements may include: acquiring market fluctuation monitoring data and adjusting the evaluation weight parameters based on the market fluctuation monitoring data; and using a multi-task Transformer model to analyze the dynamic correlation network graph based on the adjusted evaluation weight parameters to obtain the target analysis result corresponding to the target enterprise's technological achievements. The coordination risk probability is a comprehensive evaluation result, obtained by multiplying the absolute difference between the technology maturity level and the market fit index by the correlation tightness.

[0062] Furthermore, to address the insufficient market sensitivity of traditional models, achieve minute-level response to market fluctuations, and resolve the industry pain point of delayed market response caused by sudden technological advancements, this application employs a dynamic model update method combining short-term and long-term cycles. Specifically, the enterprise technology achievement evaluation method may further include: updating the preset deep learning algorithm based on a dual-cycle dynamic update mechanism; the dual-cycle dynamic update mechanism includes a target short-term cycle update and a target long-term cycle update, where the target short-term cycle update is a model update operation triggered by market fluctuation events; and the target long-term cycle update is a model update operation triggered by technological breakthrough events. That is, the short-term cycle fine-tunes the market adaptation branch parameters based on market fluctuations; the long-term cycle is triggered by major events such as technological breakthroughs, resulting in graph feature updates and full model retraining.

[0063] As can be seen from the above, the embodiments of this application access multi-source data from technical dimensions such as patents, papers, and R&D data, and market dimensions such as user behavior, industry reports, and competitive intelligence. By mapping, the dynamic offset between the technology stage and the market window period is associated, and by semantically aligning patent terms with market concepts, the problem of data fragmentation is solved. A dynamic technology and market association network is generated, with technology nodes and market nodes as entities. A cross-domain association graph is constructed through a cross-domain attention mechanism to capture the implicit relationship between technology evolution and market demand. Based on a multi-task Transformer model, cross-domain features are extracted through a shared encoder, and three indicators for intelligent evaluation are output through dynamic weight adjustment, so as to achieve joint intelligent prediction of technology maturity and market adaptability.

[0064] See Figure 5 As shown in the figure, this application discloses an enterprise technology achievement evaluation device, including:

[0065] Data acquisition module 11 is used to acquire multi-source data related to the target company's technological achievements, including technology dimension data and market dimension data;

[0066] The feature fusion module 12 is used to perform feature fusion operations on the technical dimension data and the market dimension data using a preset spatiotemporal alignment engine and a preset semantic alignment engine to obtain a fused feature vector; the technical dimension data includes patent data, paper data, and R&D data; the market dimension data includes user behavior data, industry report data, and competitive intelligence data; the fused feature vector includes a technical feature vector and a market feature vector.

[0067] The graph construction module 13 is used to construct technology nodes and market nodes based on the fused feature vectors, and to utilize a cross-domain attention network created based on a bidirectional attention mechanism to construct a dynamic association network graph based on the technology nodes and the market nodes; the dynamic association network graph is a graph representing the association between the technology nodes and the market nodes.

[0068] The result generation module 14 is used to analyze the dynamic correlation network graph using a preset deep learning algorithm to obtain the target analysis results corresponding to the technological achievements of the target enterprise; the target analysis results include technology maturity level, market adaptability index and coordination risk probability.

[0069] In some specific embodiments, the feature fusion module 12 may specifically include:

[0070] The mapping benchmark offset rule design unit is used to design mapping benchmark offset rules about the technology stage to the market window period based on the preset spatiotemporal alignment engine and dynamic offset mechanism.

[0071] The data mapping unit is used to generate a dynamic offset related to the intensity of technological breakthroughs based on a preset compensation algorithm using the preset spatiotemporal alignment engine, and to map the technology dimension data and the market dimension data to a unified timeline based on the mapping benchmark offset rule and the dynamic offset, so as to obtain spatiotemporally aligned technology dimension data and market dimension data.

[0072] The fusion feature vector generation unit is used to perform feature fusion operations on the spatiotemporally aligned technical dimension data and market dimension data using the preset semantic alignment engine in a dual-path vectorization manner to obtain a fusion feature vector.

[0073] In some specific embodiments, the map construction module 13 may specifically include:

[0074] The first node generation unit is used to deepen the technical feature vector based on the international patent classification tree to obtain technical nodes;

[0075] The second node generation unit is used to extend the market feature vector based on the industry chain to obtain market nodes;

[0076] The association construction unit is used to utilize a cross-domain attention network created based on a bidirectional attention mechanism, and to establish a first association relationship representing the contribution of technology to the market and a second association relationship representing the market's traction to technology based on the technology node and the market node, and to calculate the first association strength of the first association relationship and the second association strength of the second association relationship using a preset bidirectional attention mechanism.

[0077] The dynamic association network graph construction unit is used to construct a dynamic association network graph based on the first association relationship, the second association relationship, the technology node, and the market node.

[0078] In some specific embodiments, the result generation module 14 may specifically include:

[0079] The evaluation weight parameter adjustment unit is used to acquire market fluctuation monitoring data and adjust the evaluation weight parameters based on the market fluctuation monitoring data.

[0080] The target analysis result generation unit is used to analyze the dynamic correlation network graph based on the adjusted evaluation weight parameters using a preset deep learning algorithm to obtain the target analysis result corresponding to the target enterprise's technological achievements.

[0081] In some specific embodiments, the enterprise technology achievement evaluation device may further include:

[0082] The dynamic association network graph optimization unit is used to eliminate technical nodes and market nodes whose association tightness is lower than a preset strength threshold or whose change in association tightness within a preset time is greater than a preset change threshold, based on a preset dynamic maintenance strategy, so as to complete the optimization of the dynamic association network graph; the value of the association tightness is obtained by geometrically averaging the first association strength and the second association strength.

[0083] The deep learning algorithm update unit is used to update the preset deep learning algorithm based on a dual-loop dynamic update mechanism. The dual-loop dynamic update mechanism includes a target short-term loop update and a target long-term loop update. The target short-term loop update is a model update operation triggered by market fluctuation events, and the target long-term loop update is a model update operation triggered by technological breakthrough events.

[0084] Furthermore, embodiments of this application also disclose an electronic device, Figure 6 This is a structural diagram of an electronic device 20 according to an exemplary embodiment. The content of the diagram should not be construed as limiting the scope of this application. Specifically, the electronic device 20 may include: at least one processor 21, at least one memory 22, a power supply 23, a communication interface 24, an input / output interface 25, and a communication bus 26. The memory 22 stores a computer program, which is loaded and executed by the processor 21 to implement the relevant steps in the enterprise technology achievement evaluation method disclosed in any of the foregoing embodiments. Furthermore, the electronic device 20 in this embodiment may specifically be an electronic computer.

[0085] In this embodiment, the power supply 23 is used to provide operating voltage for each hardware device on the electronic device 20; the communication interface 24 can create a data transmission channel between the electronic device 20 and external devices, and the communication protocol it follows can be any communication protocol applicable to the technical solution of this application, and is not specifically limited here; the input / output interface 25 is used to acquire external input data or output data to the outside world, and its specific interface type can be selected according to specific application needs, and is not specifically limited here.

[0086] In addition, the memory 22, as a carrier for resource storage, can be a read-only memory, random access memory, disk or optical disk, etc. The resources stored thereon can include operating system 221, computer program 222, etc., and the storage method can be temporary storage or permanent storage.

[0087] The operating system 221 is used to manage and control the various hardware devices on the electronic device 20 and the computer program 222, which may be Windows Server, Netware, Unix, Linux, etc. In addition to including computer programs capable of performing the enterprise technology achievement evaluation method executed by the electronic device 20 as disclosed in any of the foregoing embodiments, the computer program 222 may further include computer programs capable of performing other specific tasks.

[0088] Furthermore, this application also discloses a computer-readable storage medium for storing a computer program; wherein, when the computer program is executed by a processor, it implements the aforementioned disclosed method for evaluating enterprise technological achievements. Specific steps of this method can be found in the corresponding content disclosed in the foregoing embodiments, and will not be repeated here.

[0089] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatus disclosed in the embodiments, since it corresponds to the method disclosed in the embodiments, the description is relatively simple; relevant parts can be referred to in the method section.

[0090] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0091] The steps of the methods or algorithms described in conjunction with the embodiments disclosed herein can be implemented directly by hardware, a software module executed by a processor, or a combination of both. The software module can be located in random access memory (RAM), main memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, removable disk, CD-ROM, or any other form of storage medium known in the art.

[0092] Finally, it should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0093] The technical solutions provided in this application have been described in detail above. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this application. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A method for evaluating enterprise technological achievements, characterized in that, include: Acquire multi-source data related to the target company's technological achievements, including both technology-related and market-related data. The technical dimension data and the market dimension data are fused using a preset spatiotemporal alignment engine and a preset semantic alignment engine to obtain a fused feature vector. The technical dimension data includes patent data, paper data, and R&D data; the market dimension data includes user behavior data, industry report data, and competitive intelligence data; the fused feature vector includes a technical feature vector and a market feature vector. Based on the fused feature vectors, technology nodes and market nodes are constructed, and a cross-domain attention network created based on a bidirectional attention mechanism is used to construct a dynamic relational network graph based on the technology nodes and the market nodes. The dynamic correlation network graph is a graph that represents the correlation between the technology nodes and the market nodes; The dynamic correlation network graph is analyzed using a preset deep learning algorithm to obtain the target analysis results corresponding to the technological achievements of the target enterprise; The target analysis results include technology maturity level, market fit index, and coordination risk probability; The step of using a preset spatiotemporal alignment engine and a preset semantic alignment engine to perform feature fusion operations on the technical dimension data and the market dimension data to obtain a fused feature vector includes: The preset spatiotemporal alignment engine is used to design a mapping benchmark offset rule for the transition from technology stage to market window period based on a dynamic offset mechanism; The preset spatiotemporal alignment engine generates a dynamic offset related to the intensity of technological breakthroughs based on a preset compensation algorithm. Based on the mapping benchmark offset rule and the dynamic offset, the technology dimension data and the market dimension data are mapped to a unified timeline to obtain spatiotemporally aligned technology dimension data and market dimension data. The preset semantic alignment engine is used to perform feature fusion operation on the spatiotemporally aligned technical dimension data and market dimension data in a dual-path vectorization manner to obtain a fused feature vector; The method of utilizing a cross-domain attention network created based on a bidirectional attention mechanism, and constructing a dynamic relational network graph based on the technology nodes and the market nodes, includes: A cross-domain attention network based on a bidirectional attention mechanism is used, and a first correlation relationship representing the contribution of technology to the market and a second correlation relationship representing the market's traction to technology are established based on the technology node and the market node. The first correlation strength of the first correlation relationship and the second correlation strength of the second correlation relationship are calculated using a preset bidirectional attention mechanism. A dynamic relational network graph is constructed based on the first relation, the second relation, the technology node, and the market node.

2. The enterprise technological achievement evaluation method according to claim 1, characterized in that, The construction of technical nodes and market nodes based on the fused feature vectors includes: The technical feature vector is further refined based on the international patent classification tree to obtain technical nodes; The market feature vector is extended based on the industry chain to obtain market nodes.

3. The enterprise technological achievement evaluation method according to claim 1, characterized in that, The method, which utilizes a cross-domain attention network created based on a bidirectional attention mechanism and constructs a dynamic relational network graph based on the technology nodes and the market nodes, further includes: Based on a preset dynamic maintenance strategy, technical nodes and market nodes with a correlation tightness lower than a preset strength threshold or whose correlation tightness changes more than a preset change threshold within a preset time are removed to optimize the dynamic correlation network graph; the correlation tightness is the value obtained by geometrically averaging the first correlation strength and the second correlation strength.

4. The enterprise technological achievement evaluation method according to claim 1, characterized in that, The step of analyzing the dynamic correlation network graph using a preset deep learning algorithm to obtain the target analysis results corresponding to the target enterprise's technological achievements includes: Acquire market fluctuation monitoring data and adjust the evaluation weight parameters based on the market fluctuation monitoring data; The dynamic correlation network graph is analyzed using a preset deep learning algorithm based on the adjusted evaluation weight parameters to obtain the target analysis results corresponding to the technological achievements of the target enterprise.

5. The enterprise technological achievement evaluation method according to any one of claims 1 to 4, characterized in that, Also includes: The preset deep learning algorithm is updated based on a dual-loop dynamic update mechanism; The dual-cycle dynamic update mechanism includes a short-term target cycle update and a long-term target cycle update. The short-term target cycle update is a model update operation triggered by market fluctuation events; the long-term target cycle update is a model update operation triggered by technological breakthrough events.

6. A device for evaluating enterprise technological achievements, characterized in that, include: The data acquisition module is used to acquire multi-source data related to the target company's technological achievements, including technical and market-related data. The feature fusion module is used to perform feature fusion operations on the technical dimension data and the market dimension data using a preset spatiotemporal alignment engine and a preset semantic alignment engine to obtain a fused feature vector; the technical dimension data includes patent data, paper data, and R&D data; the market dimension data includes user behavior data, industry report data, and competitive intelligence data; the fused feature vector includes a technical feature vector and a market feature vector. The graph construction module is used to construct technology nodes and market nodes based on the fused feature vectors, and to construct a dynamic association network graph based on the technology nodes and market nodes by utilizing a cross-domain attention network created based on a bidirectional attention mechanism. The dynamic correlation network graph is a graph that represents the correlation between the technology nodes and the market nodes; The result generation module is used to analyze the dynamic correlation network graph using a preset deep learning algorithm to obtain the target analysis results corresponding to the technological achievements of the target enterprise. The target analysis results include technology maturity level, market fit index, and coordination risk probability; The feature fusion module includes: The mapping benchmark offset rule design unit is used to design mapping benchmark offset rules about the technology stage to the market window period based on the preset spatiotemporal alignment engine and dynamic offset mechanism. The data mapping unit is used to generate a dynamic offset related to the intensity of technological breakthroughs based on a preset compensation algorithm using the preset spatiotemporal alignment engine, and to map the technology dimension data and the market dimension data to a unified timeline based on the mapping benchmark offset rule and the dynamic offset, so as to obtain spatiotemporally aligned technology dimension data and market dimension data. The fusion feature vector generation unit is used to perform feature fusion operations on the spatiotemporally aligned technical dimension data and market dimension data using the preset semantic alignment engine in a dual-path vectorization manner to obtain a fusion feature vector. The map construction module includes: The association construction unit is used to utilize a cross-domain attention network created based on a bidirectional attention mechanism, and to establish a first association relationship representing the contribution of technology to the market and a second association relationship representing the market's traction to technology based on the technology node and the market node, and to calculate the first association strength of the first association relationship and the second association strength of the second association relationship using a preset bidirectional attention mechanism. The dynamic association network graph construction unit is used to construct a dynamic association network graph based on the first association relationship, the second association relationship, the technology node, and the market node.

7. An electronic device, characterized in that, include: Memory, used to store computer programs; A processor for executing the computer program to implement the enterprise technology achievement evaluation method as described in any one of claims 1 to 5.

8. A computer-readable storage medium, characterized in that, Used to store a computer program, wherein the computer program, when executed by a processor, implements the enterprise technology achievement evaluation method as described in any one of claims 1 to 5.

Images