Product association feature-based intellectual property infringement risk early warning method

Abstract

The invention relates to the technical field of intellectual property protection, in particular to an intellectual property infringement risk early warning method based on product association features, which comprises the following steps: S1, identifying basic patents and key patents from an enterprise patent library; s2, constructing a right protection parameter system, wherein the parameter system comprises right protection difficulty D, evidence obtaining difficulty E, right protection income R and right protection cost C, and calculating a right protection priority index P based on the right protection parameter system; and S3, dividing risk levels according to the right protection priority index P, and outputting a corresponding right protection strategy. According to the method, key patents associated with multiple products are accurately positioned through case-product association feature analysis, case classification is realized based on association of multi-dimensional technical features, and the accuracy of case and product mapping is improved; and integrating right protection difficulty, evidence obtaining difficulty and income, calculating a benefit index through a model, and providing data support for right protection decision.

Description

Technical Field

[0001] This application relates to the field of intellectual property protection technology, and in particular to a method for early warning of intellectual property infringement risks based on product association characteristics. Background Technology

[0002] In intellectual property management, some patents (basic patents) have a wide range of technical features and may be associated with multiple products of a company. Once such patents are infringed or are infringed, the impact is often more significant. When conducting infringement risk assessments, existing technologies often treat all patents as equal objects, failing to establish differentiated assessment strategies for basic patents related to multiple products. This results in insufficient sensitivity of risk warnings for key patents, making it difficult to meet the needs of precise risk management.

[0003] In existing solutions, the association between patents and products is mostly based on manual annotation, which is inefficient and prone to omissions; case classification relies on only a single technical dimension, failing to reflect the correlation of technical features; and risk assessment models do not consider the impact of the breadth of product associations of patents on risk levels, resulting in insufficient targeted early warnings. Therefore, there is an urgent need for a method that can automatically identify basic patents and achieve differentiated risk early warnings based on case-product association characteristics. Summary of the Invention

[0004] The technical problem that this invention aims to solve is that existing technologies lack sufficient sensitivity to risk warnings for key patents.

[0005] To address this, the present invention provides a method and system for early warning of intellectual property infringement risks based on product association characteristics. This method classifies cases by constructing a technical feature association network, establishes a multi-dimensional mapping relationship between cases and enterprise products, identifies underlying patents that are simultaneously associated with multiple products, and enhances the risk assessment weight by combining the characteristics of the underlying patents, thereby achieving targeted infringement risk warnings. This invention can accurately identify the infringement and being infringed risks of key patents, improving the targeting and effectiveness of intellectual property risk management.

[0006] The technical solution adopted by this invention to solve its technical problem is: A method for early warning of intellectual property infringement risks based on product association characteristics includes the following steps: S1: Identify basic and key patents from the enterprise patent database; S2: Construct a rights protection parameter system, which includes rights protection difficulty D, evidence collection difficulty E, rights protection benefits R, rights protection costs C, and calculate the rights protection priority index P based on the rights protection parameter system; S3: Divide the risk level according to the rights protection priority index P and output the corresponding rights protection strategy.

[0007] Further, in step S1, the identification of the basic patent includes: A correlation threshold S0 is set. When the overall correlation S between a patent and at least two products is ≥ S0, it is initially determined to be a basic patent. The determination of the basic patent is supplemented and revised by technology reuse verification and historical application data.

[0008] Furthermore, the technology reuse verification includes: Check whether the patent claims include descriptions of cross-product applications; Calculate the matching degree between the patent and each product, where the matching degree = ∑(feature similarity × feature weight) / 5.

[0009] Furthermore, in step S1, the identification of the key patents includes: Extract product revenue data, profit margin, and market share from company documents; Products that account for ≥15% of the company’s total revenue or have a profit margin ≥20% of the industry average for three consecutive quarters are selected as key products. Patents with a matching degree of ≥0.7 among the patents associated with key products are marked as key patents.

[0010] Furthermore, in step S2, the calculation of the difficulty D of rights protection includes: Based on patent stability (D1), clarity of infringement determination (D2), and complexity of judicial / administrative procedures (D3), the formula is D = β. D1 ×D1 + β D2 ×D2 + β D3 Calculated using ×D3, where β D1 , β D2 , β D3 These are the weighting coefficients for D1, D2, and D3, with values ​​ranging from 0 to 1.

[0011] Furthermore, in step S2, the calculation of the evidence collection difficulty E includes: Based on the public nature of the infringement (E1), the completeness of the chain of evidence (E2), and the support of third-party evidence (E3), according to the formula E = β E1 ×E1 +β E2 ×E2 + β E3 Calculated using ×E3, where β E1 , β E2 , β E3 These are the weighting coefficients for E1, E2, and E3, with values ​​ranging from 0 to 1.

[0012] Furthermore, in step S2, the benefit R from rights protection includes the revised benefit R. 修正 The RR修正 = R ×(0.7×D1 + 0.3×E), where R is the comprehensive score calculated based on the quantification of infringement losses (R1), market competition recovery (R2), and brand protection value (R3).

[0013] Furthermore, in step S2, the calculation of the rights protection cost C includes: Based on direct cost C1, time cost C2, and opportunity cost C3, according to the formula C = β C1 ×C1 + β C2 ×C2 +β C3 Calculated using ×C3, where β C1 β C2 β C3 These are the weighting coefficients for C1, C2, and C3, with values ​​ranging from 0 to 1. When E < 0.4, the cost of rights protection C is reduced by 20%-50%.

[0014] Furthermore, in step S3, the formula for calculating the rights protection priority index P is: P = a×(1 - D) +b×(1 - E) +c×R 修正 + d×(1 - C), where a, b, c, and d are weighting coefficients.

[0015] Furthermore, in step S4, the risk level includes four levels: When P ≥ 0.8, the risk level is Level 1, and rights protection should be initiated immediately; When 0.6 ≤ P < 0.8, the risk level is level two, and after issuing a lawyer's letter warning, negotiation or litigation may be pursued. When 0.4 ≤ P < 0.6, the risk level is level three, and dynamic monitoring and patent licensing should be attempted. When P < 0.4, the risk level is level four, and one should abandon the initiative to protect one's rights and guard against counterclaims.

[0016] The beneficial effects of this invention are: 1. Automatic identification of basic patents: Through case-product association feature analysis, key patents related to multiple products can be accurately located, solving the problem of low efficiency in manual identification; 2. Differentiated risk assessment: Set higher risk weights and more sensitive early warning thresholds for basic patents to improve the risk response speed for key patents; 3. Technical Feature Association Network: Case classification is achieved based on the correlation of multi-dimensional technical features, improving the accuracy of case-product mapping; 4. Quantify the benefits of rights protection: Integrate the difficulty of rights protection, the difficulty of obtaining evidence, and the benefits, and calculate the benefit index through the model to provide data support for rights protection decisions; 5. Standardized Feature System: Unify the dimensions of technical feature descriptions for cases and products, providing a reliable foundation for correlation analysis and enhancing the system's versatility and scalability. Attached Figure Description

[0017] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0018] Figure 1 This is a schematic diagram of the method flow of the present invention. Detailed Implementation

[0019] The present invention will now be described in further detail with reference to the accompanying drawings. These drawings are simplified schematic diagrams, illustrating only the basic structure of the invention, and therefore only show the components relevant to the invention.

[0020] In the description of this invention, it should be understood that the terms "center," "longitudinal," "lateral," "length," "width," "thickness," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," "outer," "clockwise," "counterclockwise," "axial," "radial," and "circumferential," etc., indicating orientation or positional relationships, are based on the orientation or positional relationships shown in the accompanying drawings and are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the invention. Furthermore, features defined with "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this invention, unless otherwise stated, "a plurality of" means two or more.

[0021] In the description of this invention, it should be noted that, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0022] A method for early warning of intellectual property infringement risks based on product association characteristics, including... Step S1: Construction of Technical Feature Database The technical features database includes historical infringement data, patent data, and product data.

[0023] Historical infringement data: For historical infringement cases (such as patent infringement lawsuits and trademark infringement cases), natural language processing (NLP) technology is used to extract technical features from case documents (complaints and judgments), including technical fields (such as "new energy vehicles"), core functions (such as "battery management"), key components (such as "BMS chip"), technical parameters (such as "range ≥ 500km"), and application scenarios (such as "passenger vehicles"). Historical infringement data helps companies understand which technologies are prone to infringement and which technologies in products are easily updated and iterated.

[0024] Patent data includes textual data such as claims, specifications, and abstracts of patents that the company has applied for, including patent ID and descriptions of technical features (such as core functions, key components, and technical solutions). Product data includes a company's product list and product application characteristics (such as product name, function and purpose, technical parameters, and application scenarios), which can be extracted from product manuals, official website introductions, and R&D documents.

[0025] Then, the historical infringement data, patent data, and product data are normalized, including term alignment (such as merging "deep neural network" and "DNN"), synonym merging (such as merging "application" and "usage scenario"), and constructing a hierarchical technical feature tree (such as the top-level node being "technical field", the second-level node being "core function", and the third-level node being "key component").

[0026] Feature extraction is performed based on a technical feature database, and the patent text is segmented and encoded using a BERT pre-trained model to generate vector representations.

[0027] The core technical features of the patent are extracted using a keyword weighting algorithm (such as TF-IDF), forming a patent feature set T = {t1, t2, ..., t}. m}, t i 'm' represents the technical features, and 'm' represents the number of features.

[0028] S2, Patent-Product Relevance Calculation Module To accurately identify a company's core technological assets, it is necessary to first screen out basic and key patents as the core objects of subsequent rights protection assessments. This involves linking the patent protection objects with the company's actual sales products, concretizing the value of intellectual property, evaluating patent value, and assessing intellectual property risks.

[0029] S2.1 Patent-Product Relevance Score The strength of the connection between patents and products is quantified by scoring "technology matching degree + functional adaptability": S2.1.1 Technical Matching Degree (T): The cosine similarity between the keyword vector of patent technology and the keyword vector of product technology requirements is calculated using the following formula:

[0030] Among them, P i For the keyword weight of patent technology, Q i The weight of keywords related to product technology requirements is calculated using the TF-IDF algorithm, highlighting core technologies.

[0031] In this model, the weights of patent technology keywords (P) are calculated using the TF-IDF algorithm. i ) and the keyword weight of product technology requirements (Q) i The process aims to quantify the "importance" of keywords in a specific text set—considering both the frequency of keyword occurrence in individual texts and suppressing high-frequency words that appear prevalent throughout the entire text set (these words contribute less to distinguishing text features). The specific steps are as follows: TF (Term Frequency) is used to represent the frequency of a keyword in a single text, reflecting the "local importance" of the word in the current text; IDF (Inverse Document Frequency) is used to represent the logarithm of the proportion of documents in which a keyword appears in the entire text set, reflecting the "uniqueness" of the word in the global context (the fewer documents that appear, the higher the IDF value, indicating that the word has a stronger distinguishing effect on the text).

[0032] The TF-IDF weight is the product of TF and IDF, which comprehensively reflects the local frequency and global uniqueness of a keyword. The higher the value, the stronger the representativeness of the keyword to the text features.

[0033] Patent technology keyword weight P i Keyword weight Q related to product technology requirements i The calculation logic is consistent, based on the patent technology keyword weight P. i For example: First, obtain the keywords: In calculating the weight P of patent technology keywords i At that time, the text set consisted of the "claims + description" texts of all patents in the company's patent database (each patent's text was a "document"). This was used to calculate the keyword weight Q for product technical requirements. i At that time, the text set is the "Technical Requirements Description" text of all products in the enterprise's product list (the technical requirements of each product are one "document").

[0034] Preprocess each document in the text collection: Word Segmentation: Use NLP tools (such as Jieba Segmentation, spaCy) to split continuous text into independent keywords (e.g., splitting "autofocus camera" into "autofocus" and "camera"); Denoising: Filter out stop words (such as "of", "in", "for") and punctuation marks that have no practical meaning, and retain core vocabulary related to technology (such as "sensor", "algorithm", "material", etc.).

[0035] Calculate TF value: For a keyword t, the formula for the term frequency TF(t,d) in a patent document d is: TF(t,d) = The number of times keyword t appears in document d / The total number of all keywords in document d Example: If "camera" appears 5 times in patent document d and the total number of keywords in the document is 100, then TF(camera,d)=5 / 100 = 0.05.

[0036] Calculate IDF value: For a keyword t, the formula for the inverse document frequency IDF(t) in the entire patent text collection (containing N documents) is: IDF(t) = log((N + 1) / (The number of documents containing keyword t + 1)) + 1 Adding 1 in the formula is to avoid the denominator being 0 and ensure that the IDF value is positive. Example: If the patent text collection has a total of 1000 documents (N = 1000), and 10 of them contain "autofocus", then IDF(autofocus)=log((1000 + 1) / (10 + 1)) + 1) ≈ 5.51; if "sensor" appears in 800 documents, then IDF(sensor)=log(1001 / 801) + 1 ≈ 1.22. It can be seen that the IDF value of "autofocus" is higher, indicating that it is more unique globally.

[0037] Calculate TF-IDF weight (i.e., P i or Q i ) The TF-IDF weight of a keyword t in document d is: TF-IDF(t,d) = TF(t,d) × IDF(t) Example: Continuing the above example, if the TF of "autofocus" in document d is 0.03, then its TF-IDF weight = 0.03 × 5.51 ≈ 0.165; if the TF of "sensor" in document d is 0.08, then its TF-IDF weight = 0.08 × 1.22 ≈ 0.098. Obviously, the weight of "autofocus" in this document is higher and can better represent the technical features of this patent.

[0038] Through the above process, each P in the patent technology keyword vector i(i.e., the TF-IDF weight of a keyword), which can highlight core technical terms that "occur frequently and are rare in the entire patent database" in the patent text (such as proprietary technology in a specific scenario); similarly, the Q of product technical requirement keywords i It can highlight the product's unique technical requirements (such as "battery thermal management" being given more weight than "circuit design" in the context of "new energy vehicles").

[0039] This weighting method makes the subsequent calculation of "technology matching degree (T)" more accurate—similar core technology terms have higher weights and contribute more to cosine similarity, thereby improving the quantitative accuracy of the association between patents and product technologies.

[0040] S2.1.2 Functional Adaptability (F): Construct a function mapping dictionary (e.g., "reduced energy consumption" corresponds to "energy-saving product"), and calculate the overlap between patent function tags and product function requirements through text matching. Formula: F = Number of functional tags shared by the patent and the product / Total number of functional tags in the product S2.1.3 Overall Relevance (S): The weighted fusion of T and F can be adjusted according to the company's needs (e.g., increasing the weight of T for technology-driven products and increasing the weight of F for function-oriented products):

[0041] S2.2 Basic Patent Identification Basic patents refer to core technology patents that can be applied across products. The objects protected by basic patents can be applied in multiple products and often contribute significantly to the value of the products.

[0042] Patent classification and cross-identification rules: Set a correlation threshold S0 (e.g., 0.6). When the correlation between a patent and a certain product is S≧S0, the patent is determined to belong to the application patent pool of that product.

[0043] If a patent has a combined relevance of ≥S0 to at least two products, it is marked as a basic patent.

[0044] Further improve accuracy through dual verification: S2.2.1 Technology reuse verification: Check whether the patent claims explicitly state cross-product descriptions such as "applicable to XX scenario and YY scenario"; For each type of case, calculate the matching degree with each of the company's products, using the following formula: Matching degree = ∑(feature similarity × feature weight) / 5 (total number of features is 5) The feature weights are as follows: technical field 0.3, core function 0.25, key component 0.2, technical parameter 0.15, and application scenario 0.1 (determined based on the contribution of features in historical infringement determinations). S2.2.2 Supplementary Historical Application Data: The correlation calculation results are corrected by combining the company's internal patent application records (such as the patents being actually implemented in products A and B).

[0045] S2.3 Key Patent Identification Financial Report Analysis: OCR (Optical Character Recognition) technology is used to identify product revenue data in the financial report PDF, and NLP technology is combined to extract indicators such as quarterly revenue, profit margin, and market share for each product. Key product identification: Select products that account for ≥15% of the company's total revenue for three consecutive quarters (e.g., "Model X" accounts for 22% of a car company's total revenue), or have a profit margin ≥20% of the industry average as key products. Key patent marking: Patents with a matching degree of ≥0.7 in cases related to key products are marked as key patents (such as the "battery thermal management patent" related to Model X).

[0046] Step S3: Construction of the Rights Protection Parameter System Based on the relationship between patents and products, a rights protection parameter system is constructed to evaluate the difficulty of protecting selected patents, and the risk of the patent is reasonably assessed based on the difficulty of protection. S3.1 Difficulty in protecting rights (D) For the identified basic and key patents, an evaluation system is constructed from four dimensions: difficulty of rights protection, difficulty of obtaining evidence, benefits of rights protection, and costs of rights protection, to quantify the feasibility of rights protection. Patent Stability (D1): Based on the quality of the patent grant text's claims (such as the clarity of independent claims and the hierarchical completeness of dependent claims) and invalidation history (whether invalidation has been filed and the results), a score of 0 to 1 is used (1 point indicates extremely high stability, and 0 points indicates that it is easily invalidated). Clarity of Infringement Determination (D2): The complexity of the comparison between the patent claims and the technical features of the accused infringing product (such as whether there is a need for equivalent infringement determination, whether the technical features are hidden). It is quantified into 0~1 points by comparing the overlap between the patent technical keywords and the technical description of the infringing product through NLP (the higher the overlap, the clearer the determination). Judicial / Administrative Procedural Complexity (D3): Based on the judicial practice in the place where the tort occurred (such as the average trial period and the support rate of similar cases), a score is assigned according to historical data (e.g., 1 point is awarded if the trial period is ≤6 months and the support rate is ≥80%, and 0 points are awarded if the trial period is >12 months and the support rate is <50%).

[0047] Overall score for difficulty in protecting rights: D = β D1 ×D1 + β D2 ×D2 + β D3 Calculated using ×D3, where β D1 , β D2 , β D3 These are the weighting coefficients for D1, D2, and D3, respectively. The weights are adjusted based on the company's need for "priority of rights stability" or "priority of procedural efficiency." The lower the value, the easier it is to protect rights. The values ​​for D1, D2, and D3 range from 0 to 1.

[0048] S3.2 Difficulty of obtaining evidence (E) The sub-indicators reflecting the ease of obtaining evidence of infringement include: Publicity of Infringement (E1): Whether the infringing products are openly sold (e.g., can be found on e-commerce platforms) and whether the infringing technology is openly used (e.g., displayed at industrial exhibitions). The higher the publicity, the higher the score (0~1 point). Completeness of the chain of evidence (E2): Whether cross-regional evidence collection is required, whether trade secret infringement is involved (the difficulty of obtaining evidence for secret infringement is higher than that for open product infringement), and assign a value based on the operability of obtaining evidence (0~1 points); Third-party evidence support (E3): Whether there is objective evidence (such as sales data, technical appraisal reports) that can be provided by third parties such as notary offices or industry associations. The higher the support, the higher the score (0~1 points).

[0049] Overall score for difficulty in obtaining evidence: E = β E1 ×E1 +β E2 ×E2 + β E3 Calculated using ×E3, where β E1 , β E2 , β E3 These are the weighting coefficients for E1, E2, and E3, respectively. The lower the overall score for the difficulty of obtaining evidence, the greater the difficulty of obtaining evidence.

[0050] S3.3 Benefits of Rights Protection (R) The sub-indicators reflect the direct and indirect value that can be obtained after a successful rights protection action, and include: Infringement damage quantification (R1): Based on the proportion of the market share of the patented product that is squeezed by the infringing product and the sales profit of the infringing product (which can be estimated through public financial reports or industry data), it is quantified by the ratio of "estimated compensation amount / annual sales of the patented product" (0-1 points, the higher the ratio, the higher the score). Market Competition Restoration (R2): The role of successful rights protection in eliminating competitors and restoring the company's market position (higher score if the infringing product is the mainstream product in the market), 0-1 point; Brand protection value (R3): Positive impact on the company's technological reputation and the deterrent power of its patent layout (e.g., the score for core technology patent rights protection is higher than that for peripheral patents), 0-1 points.

[0051] Overall score of benefits from rights protection: R = β R1 ×R1 + β R2 × R2 + β R3 ×R3, the higher the value, the greater the benefit.

[0052] Furthermore, a correction coefficient can be added as "probability of winning similar cases," formula: R 修正 = R × (0.7 × D1 + 0.3 × E) (Patents with high stability and easy evidence collection are easier to realize actual benefits).

[0053] S3.4 Cost of Protecting Rights (C) The sub-indicators reflecting the time and financial investment in the rights protection process include: Direct costs (C1): legal fees, notary fees, appraisal fees, etc., are quantified according to the ratio of "estimated cost / estimated revenue" (1 point for ratio ≤ 20%, 0 points for ratio > 50%). Time cost (C2): The estimated time from evidence collection to judgment (1 point for a time ≤ 6 months, 0 points for a time > 18 months). Opportunity Cost (C3): The effort required for the company's R&D and market expansion during the rights protection process (1 point for core business unaffected, 0.5 points for requiring a dedicated team to follow up).

[0054] Overall score for the cost of protecting one's rights: C = β C1 ×C1 + β C2 ×C2 +β C3 Calculated using ×C3, where β C1 β C2 β C3 These are the weighting coefficients for C1, C2, and C3, respectively. The lower the overall score for the cost of protecting rights, the higher the cost.

[0055] Rights protection costs (C): For patents that are difficult to obtain evidence (E<0.4), the time cost and legal fees will increase by 20%-50%, and the corresponding C value will be reduced (e.g., the original C=0.6, the revised C=0.48).

[0056] Based on the above four dimensions, S3.5 constructs a rights protection priority index (P). Combining changes in infringement behavior (such as the addition of new infringers or expansion of the scope of infringement) and changes in patent status (such as being challenged for invalidation), the P value and warning level are adjusted in real time through an incremental update mechanism, as shown in the following formula: P = a×(1 - D) +b×(1 - E) +c×R 修正+ d×(1 - C), where a, b, c, and d are weighting coefficients.

[0057] S4 Risk Level Classification Level 1 (High Priority): P≥0.8. At this level, the patent is highly stable (D1≥0.8), easy to obtain evidence (E≥0.7), and has a success rate of ≥80%. It is recommended to initiate rights protection immediately and prioritize litigation (punitive damages can be claimed) to create a deterrent effect with the high-stability patent.

[0058] Level 2 (Medium Priority): 0.6≤P<0.8. At this level, the stability is good (D1≥0.6), the difficulty of obtaining evidence is moderate (E=0.5-0.7), and the chance of winning is 60%-80%. It is recommended to first issue a warning letter through a lawyer and simultaneously secure evidence (such as notarized purchase of infringing products). If negotiation fails, then proceed with litigation.

[0059] Level 3 (Proceed with Caution): 0.4≤P<0.6. At this point, the patent stability is generally low (D1=0.3-0.6) or it is difficult to obtain evidence (E=0.3-0.5). The chance of winning the case is 30%-60%. It is recommended to postpone litigation, prioritize monitoring infringement dynamics, try patent licensing (to make up for losses with licensing fees), and at the same time, lay out peripheral patents to enhance the bargaining power for rights protection.

[0060] Level 4 (not recommended at this time): P < 0.4 indicates poor patent stability (D1 < 0.3) or extreme difficulty in obtaining evidence (E < 0.3), with a success rate of < 30%. It is recommended to abandon proactive rights protection and focus on preventing counter-infringement (such as checking whether your own products fall under other people's patents), or replace the patented technology through technological upgrades.

[0061] Based on the above-described preferred embodiments of the present invention, and through the foregoing description, those skilled in the art can make various changes and modifications without departing from the inventive concept. The technical scope of this invention is not limited to the contents of the specification, but must be determined by the scope of the claims.

Claims

1. A method for early warning of intellectual property infringement risks based on product association characteristics, characterized in that, Includes the following steps: S1: Identify basic and key patents from the enterprise patent database; S2: Construct a rights protection parameter system, which includes rights protection difficulty D, evidence collection difficulty E, rights protection benefits R, rights protection costs C, and calculate the rights protection priority index P based on the rights protection parameter system; S3: Divide the risk level according to the rights protection priority index P and output the corresponding rights protection strategy.

2. The intellectual property infringement risk early warning method based on product association characteristics according to claim 1, characterized in that, In step S1, the identification of the basic patent includes: A correlation threshold S0 is set. When the overall correlation S between a patent and at least two products is ≥ S0, it is initially determined to be a basic patent. The determination of the basic patent is supplemented and revised by technology reuse verification and historical application data.

3. The intellectual property infringement risk early warning method based on product association characteristics according to claim 2, characterized in that, The technology reuse verification includes: Check whether the patent claims include descriptions of cross-product applications; Calculate the matching degree between the patent and each product, where the matching degree = ∑(feature similarity × feature weight) / 5.

4. The intellectual property infringement risk early warning method based on product association characteristics according to claim 1, characterized in that, In step S1, the identification of the key patents includes: Extract product revenue data, profit margin, and market share; Products that account for ≥15% of the company’s total revenue or have a profit margin ≥20% of the industry average for three consecutive quarters are selected as key products. Patents with a matching degree of ≥0.7 among the patents associated with key products are marked as key patents.

5. The intellectual property infringement risk early warning method based on product association characteristics according to claim 1, characterized in that, In step S2, the calculation of the difficulty D of rights protection includes: Based on patent stability (D1), clarity of infringement determination (D2), and complexity of judicial / administrative procedures (D3), the formula is D = β. D1 ×D1+ β D2 ×D2 + β D3 Calculated using ×D3, where β D1 , β D2 , β D3 These are the weighting coefficients for D1, D2, and D3, with values ​​ranging from 0 to 1.

6. The intellectual property infringement risk early warning method based on product association characteristics according to claim 5, characterized in that, In step S2, the calculation of the difficulty E of obtaining evidence includes: Based on the public nature of the infringement (E1), the completeness of the chain of evidence (E2), and the support of third-party evidence (E3), according to the formula E = β E1 ×E1 +β E2 ×E2 + β E3 Calculated using ×E3, where β E1 , β E2 , β E3 These are the weighting coefficients for E1, E2, and E3, with values ​​ranging from 0 to 1.

7. The intellectual property infringement risk early warning method based on product association characteristics according to claim 6, characterized in that, In step S2, the benefit R from rights protection includes the revised benefit R. 修正 The RR 修正 = R × (0.7×D1 + 0.3×E), where R is the comprehensive score calculated based on the quantification of infringement losses (R1), market competition recovery (R2), and brand protection value (R3).

8. The intellectual property infringement risk early warning method based on product association characteristics according to claim 7, characterized in that, In step S2, the calculation of the rights protection cost C includes: Based on direct cost C1, time cost C2, and opportunity cost C3, according to the formula C = β C1 ×C1 + β C2 ×C2 +β C3 Calculated using ×C3, where β C1 β C2 β C3 These are the weighting coefficients for C1, C2, and C3, with values ​​ranging from 0 to 1. When E < 0.4, the cost of rights protection C is reduced by 20%-50%.

9. The intellectual property infringement risk early warning method based on product association characteristics according to claim 8, characterized in that, In step S3, the formula for calculating the rights protection priority index P is: P = a×(1 - D) +b×(1 - E) +c×R 修正 + d×(1 - C), where a, b, c, and d are weighting coefficients.

10. The method for early warning of intellectual property infringement risks based on product association characteristics according to claim 1, characterized in that, In step S4, the risk level includes four levels: When P ≥ 0.8, the risk level is Level 1, and rights protection should be initiated immediately; When 0.6 ≤ P < 0.8, the risk level is level two, and after issuing a lawyer's letter warning, negotiation or litigation may be pursued. When 0.4 ≤ P < 0.6, the risk level is level three, and dynamic monitoring and patent licensing should be attempted. When P < 0.4, the risk level is level four, and one should abandon the initiative to protect one's rights and guard against counterclaims.

Images