A method, system and device for evaluating influence between events

Abstract

The present invention discloses a method, system and device for evaluating the influence between events. The method comprises the following steps: fusing multi-source data to construct a causal event-reasoning graph; the multi-source data at least comprises patent data and paper data; calculating the sum of the weights of the edges in the technology evolution path of the event graph by a dynamic programming method, extracting and quantifying the technology evolution path; calculating the weighted influence by the reasoning weight of the event and the weighted degree centrality, and quantifying the influence between technology events in the evolution path; the influence between events in the technology evolution path is represented in the form of a weighted matrix; the present invention quantifies the causal relationship in the technology evolution process and deeply analyzes the law of technology evolution.

Description

Technical Field

[0001] The present invention relates to a method, system and device for evaluating the influence between events. Background Art

[0002] Assessing the influence of events along a technology evolution path can provide a deeper understanding of which technological events have a critical impact on innovation and development, such as which technological events are key nodes, which events are turning points, and which events drive technological evolution. This provides valuable insights into deepening research on technology evolution paths.

[0003] Current research mainly uses patent classification, patent citations, patent texts, and the integration of multi-source data such as patents and papers to construct relationships such as common categories, co-occurrences, citations, and semantic similarities to identify technology evolution paths. However, this identification method lacks the ability to analyze technology evolution paths from a causal logic perspective, and does not analyze the impact of scientific and technological events in the path. Therefore, it is difficult to deeply understand the laws of technology evolution and cannot accurately assess the influence of events. Summary of the Invention

[0004] To solve the above technical problems, the present invention provides a method, system, and device for evaluating the influence of events, which deeply analyze the laws of technological evolution and quantify the cause-effect relationship in the process of technological evolution.

[0005] A method for evaluating the impact of events includes the following steps:

[0006] Integrate multi-source data to construct a causal event-matter map; the multi-source data includes at least patent data and paper data; the causal event-matter map includes an event map and a matter map;

[0007] Calculating the sum of the weights of the edges in the technology evolution path of the event graph by a dynamic programming method, extracting and quantifying the technology evolution path;

[0008] The weighted influence is calculated using the event's causal weight and weighted degree centrality to quantify the influence between technological events in the evolutionary path; the influence between technological events in the technological evolutionary path is represented in the form of a weighted matrix.

[0009] In some embodiments, the weighted influence is calculated by the event's event weight and weighted degree centrality to quantify the influence between technical events in the evolution path, including:

[0010] Calculate the weighted centrality of each event based on the weighted in-degree centrality and weighted out-degree centrality of events in the technology evolution path;

[0011] Combining the inter-event weights and event weighted centrality, we can get the weighted influence.

[0012] A weighted matrix is ​​constructed according to the weighted influence between events in the technology evolution path, and each element in the weighted matrix represents the weighted influence between events in the technology evolution path.

[0013] In some embodiments, the weighted influence is calculated using the event's event weight and weighted degree centrality:

[0014]

[0015] in, and The events and The value of the weighted degree centrality of For events About the event The weight of reason.

[0016] In some embodiments, the weighted centrality of each event is calculated as follows:

[0017]

[0018] in, is the weight parameter, 0< <1; For events The weighted in-degree centrality of For events The weighted out-degree centrality of .

[0019] In some embodiments, constructing a causal event-reason graph includes:

[0020] Extracting the multi-source data, screening and identifying event pairs with causal relationships, and forming a data set containing causal event pairs;

[0021] Generate an event graph based on the causal event pairs; in the event graph, nodes are events, and each edge in the event graph is a causal relationship between events;

[0022] Event clustering is generalized to construct a cause-and-effect graph; in the cause-and-effect graph, nodes are causes and each edge in the cause-and-effect graph is a causal relationship between causes and effects.

[0023] In some embodiments, calculating the sum of the weights of the edges in the technology evolution path of the event graph using a dynamic programming method includes:

[0024] Starting from the source point of each edge, perform a breadth-first traversal of the event graph along the direction of technological evolution;

[0025] Every time an event node is visited, the sum of the weights of the edges between the source point and the node is calculated and recorded as the first path strength;

[0026] Starting from the visited event node, continue breadth traversal;

[0027] During the breadth traversal process, if the event node has been visited, the strength of the currently visited path is recorded as the second path strength. The first path strength is compared with the second path strength, and the larger value of the two is taken as the current path strength.

[0028] Until all technology evolution paths are iteratively traversed.

[0029] In some embodiments, quantifying the influence between technological events in an evolutionary path includes:

[0030] According to the row order analysis of the weighted matrix, the influence of events in the technology evolution path is determined according to the size of each element in the weighted matrix.

[0031] The present invention also provides an inter-event influence assessment system, comprising:

[0032] A causal event-reason graph construction module integrates multi-source data to construct a causal event-reason graph; the multi-source data includes at least patent data and paper data; the causal event-reason graph includes an event graph and a reason graph;

[0033] A weight calculation module calculates the sum of the weights of the edges in the technology evolution path of the event graph through a dynamic programming method, extracts and quantifies the technology evolution path;

[0034] The event influence calculation module calculates the weighted influence through the event's causal weight and weighted degree centrality, and quantifies the influence between events in the evolution path; the influence between events in the technology evolution path adopts the form of a weighted matrix.

[0035] The present invention also provides an event-to-event influence assessment device, comprising:

[0036] A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, any of the above-described methods for evaluating the influence between events is implemented.

[0037] Beneficial effects:

[0038] This invention provides a method, system, and device for assessing the impact of events. By calculating the weighted impact of events and constructing a weighted matrix, this method quantifies the impact of events. This helps identify key nodes and turning points in technological development, providing a scientific basis for technological decision-making and planning. It also deeply analyzes the laws of technological evolution, quantifies causal relationships during technological evolution, and improves the accuracy and reliability of technological evolution path analysis. The application of this method can promote innovative development in the field of technology. By identifying highly influential events, technological research and development and innovation can be carried out in a more targeted manner. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 A flow chart of an inter-event influence assessment method provided by the present invention;

[0040] Figure 2 This is a flow chart of step S1 of a method for evaluating the influence of events provided by the present invention;

[0041] Figure 3 This is a flow chart of step S3 of a method for evaluating the influence of events provided by the present invention. DETAILED DESCRIPTION

[0042] The present invention will be further described in detail below with reference to the accompanying drawings.

[0043] The terms "including" and "having" and any variations thereof in the embodiments of the present invention are intended to cover non-exclusive inclusions. For example, a process, method, system, product or apparatus that includes a series of steps or modules is not necessarily limited to those steps or modules explicitly listed, but may include other steps or modules not explicitly listed or inherent to these processes, methods, products or apparatus.

[0044] like Figure 1 As shown, the present invention provides a method for evaluating the influence between events, deeply analyzes the laws of technological evolution, and quantifies the causal relationship in the process of technological evolution.

[0045] Specifically, this method includes the following steps:

[0046] Step S1, integrating multi-source data to construct a causal event-matter map; the multi-source data includes at least patent data and paper data; the causal event-matter map includes an event map and a matter map.

[0047] like Figure 2 As shown, step S1 also includes:

[0048] Step S11: extract the multi-source data, screen and identify event pairs with causal relationships, and form a data set containing causal event pairs.

[0049] Exemplarily, the method of screening and identifying causal event pairs can be implemented using template matching or a large language model. Using the Jieba word segmentation tool, with the help of a self-built scientific paper keyword database, preliminary screening of patents and scientific paper texts is performed, retaining sentences with event entities and removing sentences that have no significance for subsequent analysis. The screened sentences are used to generate a syntax tree using Stanfordparser to analyze and reveal the deep structure of the sentences. The syntactic analysis results are matched with the rule template to identify and extract event pairs with obvious causal relationships. More specifically, the syntactic analysis results can be imported into Excel, and the if statement can be applied to match the rule template. For the remaining sentences, Excel is used to screen high-frequency words, and semantic judgment is used to identify and extract events containing fuzzy causal relationships.

[0050] Alternatively, the large language model ChatGLM can be used to transform the extraction of causal events and relationships into a question-answering task. Simple and practical prompt instruction templates are designed to maximize the extraction potential of the large model, control model illusions, and achieve better performance in one-shot tasks. Technical terms are labeled as causes, and function and effect terms are labeled as effects. The technology-function / effect relationship represents a causal relationship. This labeling method ensures that the large model clearly understands the meaning of causal relationships.

[0051] Step S12: Generate an event graph based on the causal event pairs; the nodes in the event graph are events, and each edge in the graph is a causal relationship between events.

[0052] The causal event pairs extracted in step S11 are imported into a graph tool to generate an event graph. In the event graph, nodes represent events, edges represent causal relationships between events, and the weights of edges are determined by the normalized number of occurrences of the causal event pairs.

[0053] Step S13: cluster and generalize the events to construct a cause-and-effect graph; in the cause-and-effect graph, nodes are causes and each edge in the cause-and-effect graph is a causal relationship between causes and effects.

[0054] Using a method combining the Bert model and the K-means algorithm, we generalize event clustering to construct a causal graph. During the clustering process, we typically conduct several comparative experiments using different configurations. We calculate performance metrics such as precision, recall, and F1 score after clustering with different numbers of clusters to determine the optimal number of clusters. In the causal graph, each node represents a causal relationship, and each edge represents the causal relationship between causal relationships. Edge weights are calculated using the same method as in the event graph, based on the frequency of causal relationships, to reveal the strength of causal logic between different causal relationships.

[0055] Step S2: Calculate the sum of the weights of the edges in the technology evolution path of the event graph by a dynamic programming method, and extract and quantify the technology evolution path.

[0056] A dynamic programming method is used to calculate the sum of the weights of the edges in the technology evolution path to extract and quantify the technology evolution path. The dynamic programming method is specifically as follows: starting from the source point of each edge, a breadth-first traversal of the event graph is performed along the direction of technology evolution; each time an event node is visited, the sum of the weights of the edges between the source point and the node is calculated and recorded as the first path strength; starting from the visited event node, the breadth-first traversal continues; during the breadth-first traversal process, if the event node has been visited, the path strength of the current visit is recorded as the second path strength, and the first path strength is compared with the second path strength, and the larger value of the two is taken as the current path strength; until all technology evolution paths are iteratively traversed. This method extracts and quantifies technology evolution paths, making it easier to sort technology evolution paths.

[0057] Step S3: Calculate the weighted influence using the event's event weight and weighted degree centrality to quantify the influence between technical events in the evolution path; the influence between technical events in the technical evolution path is represented in the form of a weighted matrix.

[0058] like Figure 3 As shown, step S3 also includes:

[0059] Step S31: Calculate the weighted centrality of each event based on the weighted in-degree centrality and weighted out-degree centrality of the events in the technology evolution path;

[0060] Calculate the weighted centrality of each event, specifically:

[0061]

[0062] in, is the weight parameter, 0< <1; For events The weighted in-degree centrality of For events The weighted outdegree centrality of .

[0063] Weight parameter is a weight parameter between 0 and 1, Different values ​​of α result in different effects of event in-degree or out-degree on weighted degree centrality. For example, the weight parameter can be assigned values ​​of 0.3, 0.5, and 0.7. If it is found that there is no impact on event ranking, it means that in-degree and out-degree are equally important, and α is set to 0.5.

[0064] In multi-source data such as patent data and paper data, it can be understood that the out-degree is the number of events that a certain event causes. The out-degree reflects the number of causal relationships emanating from the event, that is, how many events are caused by the event. For example, if event A causes event B and event C, then the out-degree of event A is 2. The in-degree indicates the number of events caused by a certain event. The in-degree reflects the number of causal relationships to the event, that is, how many events cause the event. For example, if event D is caused by event A and event C, then the in-degree of event D is 2. Out-degree centrality and indegree centrality The out-degree and in-degree are used for calculation respectively. Among them, the event in-degree centrality is calculated by the following formula:

[0065]

[0066] m is the total number of events in the network, that is, the total number of network nodes, m-1 is the maximum possible degree of the event, The number of in-degree connections a node has.

[0067] Similarly, the event out-degree centrality is calculated using the following formula:

[0068]

[0069] The number of out-degree connections that a node has.

[0070] Out-degree centrality indicates that an event, as a "cause" in a causal chain, influences other events within the network. Therefore, a higher out-degree centrality indicates that the event has a greater impact on other events in the network, reflecting its importance as a "cause" or "driving factor." In-degree centrality indicates that an event, as an "effect" in a causal chain, is referenced or influenced by other events. A higher in-degree centrality indicates that the event is influenced by more other events.

[0071] Step S32: Combine the event weights and the weighted centrality of the events to obtain the weighted influence;

[0072] The weighted influence is

[0073]

[0074] in, and The events and The value of the weighted degree centrality of For events About the event The weight of the cause and effect is determined by referring to the strength of the node relationship in the complex system theory, which is the same as the calculation method of the weight of the causal event in the event map. According to the complex system theory proposed by scholar S. Boccaletti, the strength of the relationship between network nodes is determined by the degree of interaction and dependence between them. Therefore, in this application, it is believed that the weight of the cause and effect between events depends on the weight between the cause and effect logic associated with these events. Specifically, the weight of the cause and effect is calculated by the following formula:

[0075]

[0076] In the logic diagram, Indicates cause and effect The frequency of causal relationships in the data set reflects the specific reasoning behind the event. As a cause, a matter As a result, the number of connections is calculated. It's important to note that if two events belong to the same principle, the weight between them is set to a maximum of 1. This is because events within the same principle are closely connected, sharing the same technological foundation or goals. Their influence and connection are generally stronger than those between events from different principles. Therefore, giving events within the same principle the highest weight helps accurately reflect the close connection and importance of events in the technological evolution path.

[0077] Step S33: construct a weighted matrix according to the weighted influence between events in the technology evolution path, where each element in the weighted matrix represents the weighted influence between events in the technology evolution path.

[0078] Schematically, the weighted influence The constructed weight matrix for:

[0079]

[0080] The rows and columns in the weighted matrix represent the events in the technology evolution path. , ,…, , the elements in the matrix Representing an event About the event According to the row order analysis of the weighted matrix, the influence of events in the technology evolution path is determined according to the size of each element in the weighted matrix.

[0081] It is worth noting that the elements of the weighting matrix With elements There are unequal situations, that is, the weighted influence of event A on event B and event B on event A are different. The weighted matrix is asymmetric. This is because in the calculation of weighted influence, the weight of the matter represents and Representative events About the event ,event About the event Because the interactions between technological events are usually directional, the impact of one event on another may be greater than the reverse, so the weights of events may be asymmetric.

[0082] The quantitative results of the influence between events in the technological evolution path are presented in the form of a weighted matrix, providing an in-depth interpretation of the implicit correlation between events in the technological evolution path.

[0083] For example, path A contains five events, namely A1, A2, A3, A4, and A5. After the calculation of steps S1 to S3, the influence between any two events forms a two-dimensional weighted matrix. According to the order of the matrix rows, for example, the influence value between A1 and A2, A3, and A4 technical events is significantly larger, and the influence value between A1 and A5 is smaller. In this case, it is believed that A1 may have driven the development of A2, A3, and A4, while having a relatively small impact on A5, thereby interpreting the implicit correlation between technical events.

[0084] The following examples are used to illustrate the quantitative results of the influence of events in the technical evolution path of this application:

[0085] Table 1 Weight of affairs

[0086]

[0087] Table 2 Event weighted degree centrality

[0088]

[0089] As shown in Tables 1 and 2, these are the calculation results of event weights and event-weighted centrality.

[0090] Take the path with a larger weight "fingerprint recognition → fingerprint collection → optical equipment and technology → microscope technology and type → microscope" as an example and split it into fingerprint recognition , fingerprint collection , Optical Equipment and Technology , Microscopy Techniques and Types ,microscope Wait for 5 events.

[0091] Fingerprint recognition To fingerprint collection The weight of the cause and effect is the method and weight, such as the number 7 in Table 1, the weight of the cause and effect is 0.11. To fingerprint recognition It is number 2 in Table 1, and its weight is 0.46. To microscopy techniques and types The event number is 11 in Table 1, and the event weight is 0.1. According to the above method, the event weights between events are determined in turn.

[0092] The event weighted centrality is obtained from Table 2. The weighted influence between events is calculated based on the weighted influence, taking the event weight and the event weighted centrality as parameters. The events are regarded as rows and columns, and the weighted influence between events as elements to form a weighted matrix. .

[0093]

[0094] Through the weighting matrix It can be concluded that there are significant differences in the impact of events. "Mainly drives "fingerprint collection "Optical Equipment and Technology and Microscopy Techniques and Types The development of "microscope "The impact is relatively small. "Fingerprint collection " has a certain degree of influence on other events in the path, and has a certain degree of influence on "microscope technology and type "The impact is particularly significant." Optical equipment and technology "It plays the role of bridging the middle of the path," Microscope technology and types "Only for "fingerprint collection "Make an impact," microscope "The impact on other events in the path is relatively insignificant. Relevant researchers can find On this basis, we combine factors such as technology maturity, market demand and R&D investment to explore the formation mechanism of the influence between path events, comprehensively and accurately interpret the influence of each event on other events in the process of technological evolution, and thus understand the deep-seated reasons behind these differences.

[0095] Based on the same inventive concept, this application provides an inter-event influence assessment system, including:

[0096] A causal event-reason graph construction module integrates multi-source data to construct a causal event-reason graph; the multi-source data includes at least patent data and paper data; the causal event-reason graph includes an event graph and a reason graph;

[0097] A weight calculation module calculates the sum of the weights of the edges in the technology evolution path of the event graph through a dynamic programming method, extracts and quantifies the technology evolution path;

[0098] The event influence calculation module calculates the weighted influence through the event's causal weight and weighted degree centrality, and quantifies the influence between events in the evolution path; the influence between events in the technology evolution path adopts the form of a weighted matrix.

[0099] Based on the same inventive concept, the present application also provides an inter-event influence assessment device, comprising a memory, a processor, and a computer program stored on the memory and runnable on the processor. When the processor executes the computer program, it implements an inter-event influence assessment method as described above.

[0100] This invention provides a method, system, and device for assessing the impact of events. By quantifying the impact of events, this method helps identify key nodes and turning points in technological development, providing a scientific basis for technological decision-making and planning. It also deeply analyzes the laws of technological evolution, quantifies causal relationships during technological evolution, and improves the accuracy and reliability of technological evolution path analysis. The application of this method can promote innovative development in the technology field. By identifying highly influential events, more targeted technological research and development and innovation can be carried out.

[0101] An embodiment of the present invention discloses a computer-readable storage medium storing a computer program for electronic data exchange, wherein the computer program enables a computer to execute the described method for evaluating the influence between events.

[0102] An embodiment of the present invention discloses a computer program product, which includes a non-transitory computer-readable storage medium storing a computer program, and the computer program is operable to enable a computer to execute the described method for evaluating the influence between events.

[0103] The embodiments described above are merely illustrative. Modules described as separate components may or may not be physically separate, and components shown as modules may or may not be physical modules. They may be located in one place or distributed across multiple network modules. Some or all of the modules may be selected based on actual needs to achieve the objectives of the embodiments. Persons of ordinary skill in the art will be able to understand and implement the embodiments without inventive effort.

[0104] Through the detailed description of the above embodiments, those skilled in the art will clearly understand that each embodiment can be implemented by means of software plus a necessary general hardware platform, or of course, by means of hardware. Based on this understanding, the above technical solution, in essence, or the portion that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, including a read-only memory (ROM), a random access memory (RAM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), a one-time programmable read-only memory (OTPROM), an electronically erasable programmable read-only memory (EEPROM), a compact disc read-only memory (CD-ROM), or other optical disc storage, magnetic disk storage, magnetic tape storage, or any other computer-readable medium capable of carrying or storing data.

[0105] The above are only some embodiments of the present invention. For those skilled in the art, several modifications and improvements can be made without departing from the inventive concept of the present invention, which all fall within the scope of protection of the present invention.

Claims

1. A method for evaluating the influence of events, characterized in that: The method comprises the following steps: Integrate multi-source data to construct a causal event-matter map; the multi-source data includes at least patent data and paper data; the causal event-matter map includes an event map and a matter map; Calculating the sum of the weights of the edges in the technology evolution path of the event graph by a dynamic programming method, extracting and quantifying the technology evolution path; The weighted influence is calculated using the event's event weight and weighted degree centrality to quantify the influence between technical events in the evolution path; the influence between technical events in the technical evolution path is represented in the form of a weighted matrix; Among them, the weighted influence is calculated using the event's event weight and weighted degree centrality, and the influence between technical events in the quantified evolution path includes: According to the weighted in-degree centrality and weighted out-degree centrality of events in the technology evolution path, the weighted centrality of each event is calculated; Combining the event weights and event weighted centrality, we get the weighted influence. Constructing a weighted matrix according to the weighted influence of events in the technology evolution path, wherein each element in the weighted matrix represents the weighted influence of events in the technology evolution path; According to the row order analysis of the weighted matrix, the influence size between events in the technology evolution path is determined according to the size of each element in the weighted matrix.

2. The method for evaluating the influence of events according to claim 1, characterized in that: The weighted influence is calculated using the event's logical weight and weighted degree centrality: in, and The events and The value of the weighted degree centrality of For events About Events The weight of the matter; is the weighted influence.

3. The method for evaluating the influence of events according to claim 2, characterized in that: The weighted centrality of each event is calculated as follows: in, is the weight parameter, 0< <1; For events The weighted in-degree centrality of For events The weighted out-degree centrality of .

4. The method for evaluating the influence of events according to claim 1, characterized in that: Construct a causal event-reasoning map, including: Extracting the multi-source data, screening and identifying event pairs with causal relationships, and forming a data set containing causal event pairs; Generate an event graph according to the causal event pair; in the event graph, nodes are events, and each edge in the event graph is a causal relationship between events; Event clustering is generalized to construct a cause-and-effect graph; in the cause-and-effect graph, nodes are causes and each edge in the cause-and-effect graph is a causal relationship between causes and effects.

5. The method for evaluating the influence of events according to claim 1, characterized in that: The sum of the weights of the edges in the technical evolution path of the event graph is calculated by a dynamic programming method, including: Starting from the source point of each edge, the event graph is traversed in a breadth-first manner along the direction of technological evolution; Each time an event node is visited, the sum of the weights of the edges from the source point to the node is calculated and recorded as the first path strength; Taking the visited event node as the starting point, continue the breadth traversal; During the breadth traversal process, if the event node has been visited, the strength of the currently visited path is recorded as the second path strength, and the first path strength is compared with the second path strength, and the larger value of the two is taken as the current path strength; Until all technology evolution paths are iteratively traversed.

6. An event-to-event influence assessment system, characterized in that: include: A causal event-matter map construction module integrates multi-source data to construct a causal event-matter map; the multi-source data includes at least patent data and paper data; the causal event-matter map includes an event map and a matter map; A weight calculation module, which calculates the sum of the weights of the edges in the technology evolution path of the event graph by a dynamic programming method, and extracts and quantifies the technology evolution path; The module for calculating the influence between events calculates the weighted influence through the event weight and weighted degree centrality, and quantifies the influence between events in the evolution path; The influence between events in the technology evolution path is in the form of a weighted matrix; Among them, the weighted influence is calculated using the event's event weight and weighted degree centrality, and the influence between technical events in the quantified evolution path includes: According to the weighted in-degree centrality and weighted out-degree centrality of events in the technology evolution path, the weighted centrality of each event is calculated; Combining the event weights and event weighted centrality, we get the weighted influence. Constructing a weighted matrix according to the weighted influence of events in the technology evolution path, wherein each element in the weighted matrix represents the weighted influence of events in the technology evolution path; According to the row order analysis of the weighted matrix, the influence size between events in the technology evolution path is determined according to the size of each element in the weighted matrix.

7. An event-to-event influence assessment device, characterized in that: include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the method for evaluating the influence between events as described in any one of claims 1 to 5 when executing the computer program.

Images