Technical field subversive potential prediction method, equipment and program product

Abstract

The invention discloses a technical field subversive potential prediction method and device and a program product, and the method comprises the steps: firstly obtaining patent data of a target technical field, and constructing an IPC classification set for the patent data in each preset time window; constructing a patent technology classification bipartite graph according to the association relationship between the patents in the classification set and the IPC classification numbers; based on the co-occurrence strength of any two technology classification nodes in the bipartite graph, generating a technology classification co-occurrence network; performing community division on technical classification nodes in the technical classification co-occurrence network; performing cross-time window community similarity calculation on community division results in different time windows, and constructing a community evolution event list; and finally, obtaining a subversive potential index of the target technical field according to the community evolution event list, and obtaining the subversive potential index of the target technical field based on the subversive potential index. According to the embodiment of the invention, the accuracy of subversive potential prediction in the technical field is effectively improved.

Description

Technical Field

[0001] This invention relates to the field of data mining technology, and in particular to a disruptive potential prediction method, device, and program product. Background Technology

[0002] In today's rapidly evolving technological era, disruptive innovation is crucial for the survival and development of enterprises. Accurately assessing the disruptive potential of technologies can provide important guidance for R&D decisions, strategic planning, and intellectual property management for enterprises and research institutions. Existing methods for assessing disruptive potential primarily rely on expert judgment or simple patent indicator analysis, which suffers from problems such as strong subjectivity, insufficient accuracy, and difficulty in comprehensively reflecting the complex relationships within the technological field.

[0003] Therefore, there is an urgent need to invent a method that can objectively, accurately, and comprehensively assess the potential for technological disruption, in order to solve the problems that existing methods for predicting the potential for technological disruption rely too much on expert experience, lack accuracy, and fail to fully reflect the complex relationships in the technological field. Summary of the Invention

[0004] In view of this, embodiments of the present invention provide a method, device, and program product for predicting disruptive potential in the technical field, which at least partially solves the problems existing in the prior art.

[0005] Other features and advantages of the invention will become apparent from the following detailed description, or may be learned in part by practice of the invention.

[0006] To achieve the above objectives, the embodiments of the present invention provide the following technical solutions:

[0007] According to a first aspect of the present invention, a method for predicting disruptive potential in the technical field is provided, the method comprising:

[0008] Obtain patent data to be analyzed in the target technology field;

[0009] For the patent data to be analyzed within each preset time window, construct the corresponding IPC classification set;

[0010] Based on the relationship between patents and IPC classification numbers in the IPC classification set, a bipartite graph of patent technology classification is constructed.

[0011] Based on the co-occurrence strength of any two technology classification nodes in the patent technology classification bipartite graph, a technology classification co-occurrence network is generated;

[0012] The community division of each technology category node in the technology category co-occurrence network is performed to obtain the community division results;

[0013] The community similarity of the community division results in different time windows is calculated across time windows, and a list of community evolution events is constructed based on the community similarity calculation results;

[0014] Based on the community evolution event list, a disruptive potential index for the target technology field is obtained. Based on the disruptive potential index and a weighted adaptive algorithm, a disruptive potential index for the target technology field is obtained.

[0015] Furthermore, for the patent data to be analyzed within each preset time window, a corresponding IPC classification set is constructed, including:

[0016] For the patent data to be analyzed within each preset time window, the patent data to be analyzed is preprocessed to remove patent data that lacks key information, and a set of valid patents after preprocessing is obtained. The key information includes IPC classification number and citation relationship.

[0017] The IPC classification numbers of each patent in the valid patent set are standardized to construct the corresponding standard IPC classification set.

[0018] Furthermore, based on the association between patents and IPC classification numbers in the IPC classification set, a patent technology classification bipartite graph is constructed, including:

[0019] Based on the association between patents and IPC classification numbers in the IPC classification set, a patent technology classification bipartite graph G = (P', N, E) is constructed, where P' is the patent node corresponding to the valid patent set, N is the technology classification node corresponding to the IPC classification set, and E is the edge connecting the corresponding patent node and technology classification node.

[0020] Furthermore, based on the co-occurrence strength of any two technology classification nodes in the patent technology classification bipartite graph, a technology classification co-occurrence network is generated, including:

[0021] By traversing all connections between patent nodes and technology classification nodes in the patent technology classification bipartite graph, and counting the number of times any two technology classification nodes co-occur in different patents, the co-occurrence strength between any two technology classification nodes is calculated. The calculation formula is as follows: Among them, W ij For any two technology classification nodes ni and n j The co-occurrence intensity between them, P i ∩P j For any two technology classification nodes n i and n j The number of times they appear together, P i ∪P j For any two technology classification nodes n iand n j Total number of occurrences;

[0022] The co-occurrence strength between any two technology classification nodes is used as the edge weight to generate the technology classification co-occurrence network G. N .

[0023] Furthermore, community partitioning is performed on each technology category node in the aforementioned technology category co-occurrence network to obtain the community partitioning results, including:

[0024] Using the technology classification co-occurrence network as input data, the Leiden algorithm is used to find the set of technology classification nodes with tight internal connections and sparse external connections in the technology classification co-occurrence network to form communities. The community division result is obtained through continuous iterative optimization.

[0025] Furthermore, cross-time-window community similarity calculations are performed on the community segmentation results within different time windows. Based on the community similarity calculation results, a list of community evolution events is constructed, including:

[0026] Based on the community division results of adjacent time windows, the community similarity between any two communities in adjacent time windows is calculated using the following formula: in, Let C be the set of communities corresponding to time window t. t The community in the middle, Let C be the community set corresponding to time window t+1. t+1 The community in the middle, for and Community similarity between them;

[0027] If community similarity Greater than or equal to the first preset similarity threshold, and Not with community group C t Matching other communities in, then in and Establish a one-to-one matching relationship between them;

[0028] If the community With community collection C t+1 If the community similarity between communities of a size greater than or equal to a first preset size is greater than or equal to a first preset similarity threshold, then the community evolution event between adjacent time windows is a splitting event.

[0029] If the community With community collection C t If the community similarity between communities of a size greater than or equal to the second preset size is greater than or equal to the first preset similarity threshold, then the community evolution event between adjacent time windows is a merging event.

[0030] If community similarity If the community size change rate between adjacent time windows is greater than or equal to the second preset similarity threshold, and the community size change rate between adjacent time windows is less than or equal to the preset change rate threshold, then the community evolution event between adjacent time windows is a stable event.

[0031] Based on the community evolution events between each adjacent time window, a community evolution event list is constructed. The community evolution event list includes the evolution event type, event occurrence time, community ID, node set, community size, community modularity, and core nodes.

[0032] Further, based on the community evolution event list, a disruptive potential indicator for the target technology field is obtained. Based on the disruptive potential indicator and a weighted adaptive algorithm, a disruptive potential index for the target technology field is obtained, including:

[0033] Based on the aforementioned list of community evolution events, establish disruptive potential indicators for the target technology field;

[0034] The disruptive potential indicators are standardized to obtain standardized values.

[0035] The combined weights of the standardized values ​​of each indicator are obtained by combining the entropy weight method and expert weights.

[0036] The disruptive potential index of the target technology field is calculated based on the standardized values ​​of each indicator and their corresponding comprehensive weights.

[0037] Furthermore, the disruptive potential indicators include structural stability indicators, dynamic evolution indicators, and cross-border innovation indicators;

[0038] The structural stability indicators include modularity Q and core node ratio KCP;

[0039] The modularity Q is the modularity value in the community partitioning result obtained using the Leiden algorithm;

[0040] The core nodes in the core node ratio KCP are patent nodes whose node degree centrality is greater than a preset core node screening threshold. The formula for calculating node degree centrality is: Among them, DC i For the degree centrality of patent node i, k i Let k be the degree of patent node i, representing the total number of nodes in the technology category involved in the patent. max This represents the maximum degree of a node within the community.

[0041] The dynamic evolution indicators include the cross-community citation growth rate (CRG) and the community size change rate (ΔS).

[0042] The cross-community reference growth rate (CRG) is the growth rate of the sum of edge weights referencing each other between communities within adjacent time windows, and is calculated using the following formula: in, This represents the sum of cross-community edge weights over time window t, where the cross-community edge weights are from the technology classification co-occurrence network. The edge weights corresponding to edges whose midpoints belong to different communities;

[0043] The community size change rate ΔS is the community size change rate between adjacent time windows;

[0044] The cross-disciplinary innovation index includes the Technology Crossover Index (CTI), and the formula for calculating the Technology Crossover Index (CTI) is as follows: Where, ∑n i ∈C, This indicates that there are edge connections between nodes n representing different technical categories within a community in the technical category co-occurrence network. i and community-external technology classification nodes n j The number of edge weights between them, |C|max w ij This represents the total number of edge weights within the community.

[0045] According to a second aspect of the present invention, an apparatus is provided, the apparatus comprising: a processor and a memory;

[0046] The memory is used to store one or more program instructions;

[0047] The processor is configured to run one or more program instructions to perform the steps of a method for predicting disruptive potential in the technical field as described in any of the preceding claims.

[0048] According to a third aspect of the present invention, a computer program product is provided, the computer program product comprising a computing program stored on a non-transitory computer-readable storage medium, the computer program comprising program instructions that, when executed by a computer, cause the computer to perform the steps of a method for predicting disruptive potential in the technical field as described in any of the preceding claims.

[0049] This invention provides a method, device, and program product for predicting disruptive potential in a technical field, comprising: first, acquiring patent data in the target technical field; constructing an IPC classification set for the patent data within each preset time window; then, constructing a patent technology classification bipartite graph based on the association between patents and IPC classification numbers in the classification set; next, generating a technology classification co-occurrence network based on the co-occurrence strength of any two technology classification nodes in the bipartite graph; performing community division on the technology classification nodes in the technology classification co-occurrence network; calculating cross-time window community similarity for the community division results within different time windows, and constructing a community evolution event list; finally, obtaining a disruptive potential index for the target technical field based on the community evolution event list, and obtaining a disruptive potential index for the target technical field based on the disruptive potential index. This invention effectively improves the accuracy of predicting disruptive potential in a technical field. Attached Figure Description

[0050] To more clearly illustrate the embodiments of the present invention or the technical solutions in 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 merely exemplary, and those skilled in the art can derive other embodiments based on the provided drawings without creative effort.

[0051] Figure 1 This is a flowchart illustrating a method for predicting disruptive potential in the technical field, provided as an embodiment of the present invention. Detailed Implementation

[0052] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. 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 should fall within the scope of protection of the present invention.

[0053] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0054] Figure 1 The figure illustrates a flowchart of a method for predicting disruptive potential in the technical field according to an embodiment of the present invention.

[0055] like Figure 1 As shown, the method for predicting disruptive potential in the technical field according to an embodiment of the present invention may include steps S100, S200, S300, S400, S500, S600 and S700.

[0056] In step S100, patent data to be analyzed in the target technical field is obtained.

[0057] Specifically, the above steps include:

[0058] The system allows users to input the technical field to be analyzed, and retrieves relevant patent data (including IPC classification numbers and citation relationships) from a publicly available patent database.

[0059] Next, in step S200, a corresponding IPC classification set is constructed for the patent data to be analyzed within each preset time window.

[0060] Specifically, the above steps include:

[0061] First, for the patent data to be analyzed within each preset time window (e.g., 5 years), data preprocessing is performed to remove patent data lacking key information and retain valid patents, resulting in a preprocessed set of valid patents. Key information includes IPC classification numbers and their references.

[0062] The IPC classification numbers of each patent in the valid patent set are standardized to unify the format and eliminate ambiguity, thus constructing a corresponding standard IPC classification set N = {n1, n2, ..., n}. k}

[0063] In step S300, a patent technology classification bipartite graph is constructed based on the association between patents in the IPC classification set and IPC classification numbers.

[0064] Specifically, the above steps include:

[0065] Based on the relationship between patents and IPC classification numbers in the IPC classification set, a bipartite graph of patent technology classification is constructed as G = (P', N, E), where P' is the patent node corresponding to the valid patent set, N is the technology classification node corresponding to the IPC classification set (e.g., G06F is the computer field), and E is the edge connecting the corresponding patent node and technology classification node.

[0066] Next, in step S400, a technology classification co-occurrence network is generated based on the co-occurrence strength of any two technology classification nodes in the patent technology classification bipartite graph.

[0067] Specifically, the above steps include:

[0068] After constructing the patent technology classification bipartite graph, a transformation process is performed to delve deeper into the intrinsic relationships between technology classifications. The connections between all patent nodes and technology classification nodes in the bipartite graph are traversed. By statistically analyzing the number of times any two technology classification nodes co-occur in different patents and combining this with the total number of node occurrences, the co-occurrence strength between any two technology classification nodes is calculated using the Jaccard coefficient. The calculation formula is as follows: Among them, W ij For any two technology classification nodes n i and n j The co-occurrence intensity between them, P i ∩P j For any two technology classification nodes n i and n j The number of times they appear together, P i ∪P j For any two technology classification nodes n i and n j Total number of occurrences, P i ={p∈P'|t∈N p}

[0069] The co-occurrence strength between any two technology classification nodes is used as the edge weight to generate the technology classification co-occurrence network G. N = (N, W).

[0070] Next, in step S500, the communities of each technology category node in the technology category co-occurrence network are divided to obtain the community division results.

[0071] Specifically, the above steps include:

[0072] After completing the technical classification co-occurrence network G N After setting the network to (N, W), the technology classification co-occurrence network is used as input to the Leiden algorithm. The Leiden algorithm iteratively optimizes the grouping of nodes in the network, assigning technology classification nodes with close connections to the same community. During the algorithm's operation, by maximizing metrics such as modularity, it seeks sets of nodes with tightly connected internal connections and sparse external connections, thus forming different communities. Finally, each technology classification node is assigned to its corresponding community, resulting in the community partitioning result C.

[0073] In step S600, cross-time window community similarity is calculated for the community division results within different time windows, and a community evolution event list is constructed based on the community similarity calculation results.

[0074] Specifically, the above steps include:

[0075] Co-occurrence networks are classified according to the temporal techniques corresponding to each preset time window. And the corresponding community division results C1, C2, ..., C t ,in, C represents the co-occurrence network of technology categories corresponding to the time window at time t. t This represents the community division result corresponding to the time window at time t.

[0076] Based on the community division results of adjacent time windows t and t+1, C t and C t+1 The Jaccard coefficient is used to calculate the community similarity between any two communities in adjacent time windows. The calculation formula is as follows: in, This is used to measure the proportion of technology category nodes shared by two communities. The higher the value, the stronger the technology correlation. The value range is [0,1]. Let C be the set of communities corresponding to time window t. t The community in the middle, Let C be the community set corresponding to time window t+1. t+1 The community in the middle, for and The similarity between communities.

[0077] Community correspondence establishment:

[0078] If community similarity Greater than or equal to the first preset similarity threshold θ S =0.2, and Not with community group C t If it matches other communities in the middle, then establish A one-to-one matching relationship.

[0079] If the community With community collection C t+1 If the community similarity between communities of a size greater than or equal to a first preset size is greater than or equal to a first preset similarity threshold, then the community evolution event between adjacent time windows is a splitting event. The aforementioned first preset size is...

[0080] If the community With community collection C t If the community similarity between communities of a size greater than or equal to the second preset size is greater than or equal to the first preset similarity threshold, then the community evolution event between adjacent time windows is a merging event, and the aforementioned second preset size is 1.5 times the original largest community.

[0081] If community similarity If the community size change rate between adjacent time windows is greater than or equal to the second preset similarity threshold of 0.8, and the change rate between adjacent time windows is less than or equal to the preset change rate threshold of 20%, then the community evolution event between adjacent time windows is a stable event.

[0082] Based on the community evolution events between each adjacent time window, a list of community evolution events is constructed. The list of community evolution events includes the type of evolution event, the time of occurrence of the event, the community ID, the set of nodes, the community size, the community modularity, and the core nodes.

[0083] Finally, in step S700, based on the list of community evolution events, a disruptive potential index for the target technology field is obtained, and based on the disruptive potential index and the weight adaptive algorithm, a disruptive potential index for the target technology field is obtained.

[0084] Specifically, the above steps include:

[0085] Based on the community segmentation results, the technology classification co-occurrence network, and the community evolution event list, a disruptive potential indicator system for the target technology field is established.

[0086] The aforementioned disruptive potential indicators include structural stability indicators, dynamic evolution indicators, and cross-border innovation indicators.

[0087] Structural stability indicators include modularity (Q) and core node ratio (KCP).

[0088] The modularity Q is the modularity value in the community partitioning result obtained directly from the Leiden algorithm. The value ranges from [0.5, 1]. The higher the value, the better the quality of the community partitioning and the clearer the structure.

[0089] Core node ratio The core nodes are patent nodes whose degree centrality is greater than a preset core node screening threshold. The formula for calculating the degree centrality is: Among them, DC i For the degree centrality of patent node i, k i Let k be the degree of patent node i, representing the total number of nodes in the technology category involved in the patent. max This represents the maximum degree of nodes within the community. The core node selection threshold is set to 0.8. If DC... i If the value is ≥0.8, then the patent node i is a core node.

[0090] Dynamic evolution indicators include cross-community citation growth rate (CRG) and community size change rate (ΔS).

[0091] Cross-community citation growth rate (CRG) is the growth rate of the sum of edge weights of mutual citations between communities within adjacent time windows. It reflects the diffusion speed and integration trend of technology between different communities. The higher the CRG value, the more active the communication and penetration of technology between communities, and the greater the possibility of technological disruption. Its calculation formula is as follows: in, This represents the sum of cross-community edge weights within time window t, where the cross-community edge weights are part of the technology classification co-occurrence network. The edge weights corresponding to edges whose midpoints belong to different communities (based on the community partitioning result C) t judge).

[0092] The community size change rate ΔS is the rate of change in community size between adjacent time windows.

[0093] Cross-disciplinary innovation indicators include the Technology Crossover Index (CTI), which measures the degree of connection between a technology community and other fields. A higher CTI value indicates that the community's technology is more likely to trigger disruptive innovation through cross-disciplinary integration. The calculation formula is as follows: Where, ∑n i ∈C, This indicates that there are edge connections between nodes n representing different technical categories within a community in the technical category co-occurrence network. i and community-external technology classification nodes n j The number of edge weights between them, i.e., the number of technical category nodes n traversed within the community. i and community-external technology classification nodes n j If n i With n j In a co-occurrence network, there exists an edge (w) ij If >0), then accumulate w. ij ,|C|maxw ij This represents the total number of edge weights within the community.

[0094] The dimensions and value ranges of different indicators vary greatly (for example, the community size may range from dozens to hundreds, and the modularity ranges from -1 to 1). Therefore, it is necessary to eliminate the influence of dimensions through standardization and standardize the indicators in the disruptive potential index system (for example, Min-Max normalization, and the standardized value ), and obtain the standardized values of indicators with a value range of 0-1.

[0095] Adopt a weight adaptive algorithm that combines the entropy weight method and expert weights. The entropy weight method calculates objective weights to reduce subjective influence. If Y is the standardized value of indicator X, the entropy value calculation formula is as follows: The calculation formula for entropy weight is:

[0096] Integrate expert experience weights and obtain expert weights through the AHP method

[0097] Generate comprehensive weights using entropy weights and patent weights

[0098] According to the standardized values of each indicator and their corresponding comprehensive weights, calculate the disruptive potential index of the target technical field Among them, w Q + w KCP + w ΔS + w CRG + w CTI = 1.

[0099] By setting thresholds, determine the disruptive potential of the target technical field. For example: TPI > 0.7 indicates high disruptive potential, 0.5 < TPI ≤ 0.7 indicates medium disruptive potential, and TPI ≤ 0.5 indicates low disruptive potential.

[0100] In the embodiment of the present invention, the disruptive potential of technology (TPI, Technological Disruption Potential Index) is quantified by weighted summation of multiple dimensions of indicators in the table. Its essence is to convert indicators of different natures into values on a unified scale, and finally use a comprehensive index to reflect the disruptive potential of the technical field.

[0101] In addition, the embodiment of the present invention also provides a device, which includes: a processor and a memory; the memory is used to store one or more program instructions; the processor is used to run one or more program instructions to execute the steps of a method for predicting the disruptive potential of a technical field as described above.

[0102] In addition, embodiments of the present invention also provide a computer program product, which includes computer program instructions that, when executed by a processor, implement the steps of a method for predicting disruptive potential in the technical field as described above.

[0103] The invention discloses a method, device, and program product for predicting disruptive potential in the technical field. This method integrates patent-technology classification bipartite graph modeling, dynamic community detection, and multi-dimensional indicator fusion to predict and evaluate disruptive potential. It is applicable to the early identification and dynamic monitoring of disruptive technologies across all technical fields and can be applied to scenarios such as science and technology innovation management, technology forecasting, science and technology intelligence analysis, and industrial policy formulation.

[0104] This invention achieves a two-way mapping between micro-patent data and macro-technical classification by constructing a patent technology classification bipartite graph.

[0105] This invention, based on the Leiden algorithm and time window slicing, achieves accurate capture of split / merge / stabilization events in the technology community.

[0106] The embodiments of the present invention use the entropy weight method combined with expert weights to evaluate the disruptive potential of technology. By combining subjective and objective weights, the influence of human bias is effectively reduced.

[0107] In this embodiment of the invention, the processor can be an integrated circuit chip with signal processing capabilities. The processor can be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. It can implement or execute the methods, steps, and logic block diagrams disclosed in this embodiment of the invention. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the methods disclosed in this embodiment of the invention can be directly implemented by a hardware decoding processor, or implemented by a combination of hardware and software modules in the decoding processor. The software modules can be located in random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, or other mature storage media in the art. The processor reads information from the storage medium and, in conjunction with its hardware, completes the steps of the above methods. The storage medium can be memory, for example, volatile memory or non-volatile memory, or may include both volatile and non-volatile memory. Non-volatile memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. Volatile memory can be random access memory (RAM), which is used as an external cache. By way of example, but not limitation, many forms of RAM are available, such as Static Random Access Memory (SRAM), Dynamic Random Access Memory (DRAM), Synchronous DRAM (SDRAM), Double Data Rate SDRAM (DDRSDRAM), Enhanced Synchronous DRAM (ESDRAM), Synchlink DRAM (SLDRAM), and Direct Rambus RAM (DRRAM).The storage media described in the embodiments of this invention are intended to include, but are not limited to, these and any other suitable types of memory. Those skilled in the art will recognize that the functions described in one or more of the above examples can be implemented using a combination of hardware and software. When applied software, the corresponding functions can be stored in a computer-readable medium or transmitted as one or more instructions or code on a computer-readable medium. Computer-readable media include computer storage media and communication media, wherein communication media include any medium that facilitates the transmission of computer programs from one place to another. Storage media can be any available medium accessible to general-purpose or special-purpose computers. Although the invention has been described in detail above with general description and specific embodiments, modifications or improvements can be made to it, which will be apparent to those skilled in the art. Therefore, all such modifications or improvements made without departing from the spirit of the invention are within the scope of protection claimed by this invention.

[0108] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Any simple modifications, equivalent changes, or alterations made by those skilled in the art using the disclosed technical content shall fall within the protection scope of the present invention.

Claims

1. A method for predicting disruptive potential in a technological field, characterized in that, The method includes: Obtain patent data to be analyzed in the target technology field; For the patent data to be analyzed within each preset time window, construct the corresponding IPC classification set; Based on the relationship between patents and IPC classification numbers in the IPC classification set, a bipartite graph of patent technology classification is constructed. Based on the co-occurrence strength of any two technology classification nodes in the patent technology classification bipartite graph, a technology classification co-occurrence network is generated; The community division of each technology category node in the technology category co-occurrence network is performed to obtain the community division results; The community similarity of the community division results in different time windows is calculated across time windows, and a list of community evolution events is constructed based on the community similarity calculation results; Based on the community evolution event list, a disruptive potential index for the target technology field is obtained. Based on the disruptive potential index and a weighted adaptive algorithm, a disruptive potential index for the target technology field is obtained.

2. The method for predicting disruptive potential in the technical field according to claim 1, characterized in that, For the patent data to be analyzed within each preset time window, a corresponding IPC classification set is constructed, including: For the patent data to be analyzed within each preset time window, the patent data to be analyzed is preprocessed to remove patent data that lacks key information, and a set of valid patents after preprocessing is obtained. The key information includes IPC classification number and citation relationship. The IPC classification numbers of each patent in the valid patent set are standardized to construct the corresponding standard IPC classification set.

3. The method for predicting disruptive potential in the technical field according to claim 1, characterized in that, Based on the correlation between patents and IPC classification numbers in the aforementioned IPC classification set, a bipartite graph of patent technology classification is constructed, including: Based on the association between patents and IPC classification numbers in the IPC classification set, a patent technology classification bipartite graph G = (P', N, E) is constructed, where P' is the patent node corresponding to the valid patent set, N is the technology classification node corresponding to the IPC classification set, and E is the edge connecting the corresponding patent node and technology classification node.

4. The method for predicting disruptive potential in the technical field according to claim 1, characterized in that, Based on the co-occurrence strength of any two technology classification nodes in the patent technology classification bipartite graph, a technology classification co-occurrence network is generated, including: By traversing all connections between patent nodes and technology classification nodes in the patent technology classification bipartite graph, and counting the number of times any two technology classification nodes co-occur in different patents, the co-occurrence strength between any two technology classification nodes is calculated. The calculation formula is as follows: Among them, W ij For any two technology classification nodes n i and n j The co-occurrence intensity between them, P i ∩P j For any two technology classification nodes n i and n j The number of times they appear together, P i ∪P j For any two technology classification nodes n i and n j Total number of occurrences; The co-occurrence strength between any two technology classification nodes is used as the edge weight to generate the technology classification co-occurrence network G. N .

5. The method for predicting disruptive potential in the technical field according to claim 1, characterized in that, The community division of each technology category node in the technology category co-occurrence network is performed to obtain the community division results, including: Using the technology classification co-occurrence network as input data, the Leiden algorithm is used to find the set of technology classification nodes with tight internal connections and sparse external connections in the technology classification co-occurrence network to form communities. The community division result is obtained through continuous iterative optimization.

6. The method for predicting disruptive potential in the technical field according to claim 1, characterized in that, Cross-time-window community similarity calculations are performed on the community segmentation results within different time windows. Based on the community similarity calculation results, a list of community evolution events is constructed, including: Based on the community division results of adjacent time windows, the community similarity between any two communities in adjacent time windows is calculated using the following formula: in, Let C be the set of communities corresponding to time window t. t The community in the middle, Let C be the community set corresponding to time window t+1. t+1 The community in the middle, for and Community similarity between them; If community similarity Greater than or equal to the first preset similarity threshold, and Not with community group C t Matching other communities in, then in and Establish a one-to-one matching relationship between them; If the community With community collection C t+1 If the community similarity between communities of a size greater than or equal to a first preset size is greater than or equal to a first preset similarity threshold, then the community evolution event between adjacent time windows is a splitting event. If the community With community collection C t If the community similarity between communities of a size greater than or equal to the second preset size is greater than or equal to the first preset similarity threshold, then the community evolution event between adjacent time windows is a merging event. If community similarity If the community size change rate between adjacent time windows is greater than or equal to the second preset similarity threshold, and the community size change rate between adjacent time windows is less than or equal to the preset change rate threshold, then the community evolution event between adjacent time windows is a stable event. Based on the community evolution events between each adjacent time window, a community evolution event list is constructed. The community evolution event list includes the evolution event type, event occurrence time, community ID, node set, community size, community modularity, and core nodes.

7. The method for predicting disruptive potential in the technical field according to claim 1, characterized in that, Based on the community evolution event list, a disruptive potential index for the target technology field is obtained. Based on the disruptive potential index and a weighted adaptive algorithm, a disruptive potential index for the target technology field is obtained, including: Based on the aforementioned list of community evolution events, establish disruptive potential indicators for the target technology field; The disruptive potential indicators are standardized to obtain standardized values. The combined weights of the standardized values ​​of each indicator are obtained by combining the entropy weight method and expert weights. The disruptive potential index of the target technology field is calculated based on the standardized values ​​of each indicator and their corresponding comprehensive weights.

8. The method for predicting disruptive potential in the technical field according to claim 1, characterized in that, The disruptive potential indicators include structural stability indicators, dynamic evolution indicators, and cross-border innovation indicators; The structural stability indicators include modularity Q and core node ratio KCP; The modularity Q is the modularity value in the community partitioning result obtained using the Leiden algorithm; The core nodes in the core node ratio KCP are patent nodes whose node degree centrality is greater than a preset core node screening threshold. The formula for calculating node degree centrality is: Among them, DC i For the degree centrality of patent node i, k i Let k be the degree of patent node i, representing the total number of nodes in the technology category involved in the patent. max This represents the maximum degree of a node within the community. The dynamic evolution indicators include the cross-community citation growth rate (CRG) and the community size change rate (ΔS). The cross-community reference growth rate (CRG) is the growth rate of the sum of edge weights referencing each other between communities within adjacent time windows, and is calculated using the following formula: in, This represents the sum of cross-community edge weights over time window t, where the cross-community edge weights are from the technology classification co-occurrence network. The edge weights corresponding to edges whose midpoints belong to different communities; The community size change rate ΔS is the community size change rate between adjacent time windows; The cross-disciplinary innovation index includes the Technology Crossover Index (CTI), and the formula for calculating the Technology Crossover Index (CTI) is as follows: Where, ∑n i ∈C, This indicates that there are edge connections between nodes n representing different technical categories within a community in the technical category co-occurrence network. i and community-external technology classification nodes n j The number of edge weights between them, |C|max w ij This represents the total number of edge weights within the community.

9. A device for predicting disruptive potential in the technical field, characterized in that, The device includes: a processor and a memory; The memory is used to store one or more program instructions; The processor is configured to run one or more program instructions to perform the steps of a method for predicting disruptive potential in the technical field as described in any one of claims 1 to 8.

10. A computer program product, characterized in that, The computer program product includes computer program instructions that, when executed by a processor, implement the steps of a method for predicting disruptive potential in the technical field as described in any one of claims 1 to 8.

Images