WO2019184176A1 - Method for evaluating patent value, method for constructing patent value evaluation model, and readable storage medium - Google Patents

Abstract

A method for evaluating patent value, and a method for constructing a patent value evaluation model, comprising: selecting parameters influencing patent value (S101); selecting a patent data sample according to preset patent value evaluation standards (S102); and establishing an evaluation formula reflecting the relationship between the selected parameters and the patent value on the basis of the selected patent data sample (S104). Also provided is a computer-readable medium that performs the method for evaluating patent value and the method for constructing a patent value evaluation model. With the present invention, an objective and accurate patent value evaluation model may be generated in a targeted manner, and the accuracy of patent value evaluation may be improved.

Description

Patent value evaluation method, model construction method thereof and readable storage medium

The present application claims the priority of the Chinese Patent Application No. 2018102771609, No. 2018102766526, and No. 2018102965110, filed on March 30, 2018, the entire specification, the The way it is cited in this application.

Technical field

The present disclosure relates to the field of patent analysis technology, and in particular, to a method for constructing a patent value evaluation model, a patent value evaluation method, and a computer readable storage medium.

Background technique

Chinese researchers have done a lot of research on the evaluation indicators and models of high-value patents in China. The basic idea is to construct an index system for patent value evaluation based on the three major factors affecting patent value: technology, market and law. The evaluation model calculation method mainly adopts expert scoring method, hierarchical method, decision tree method and fuzzy comprehensive evaluation method. Wan Xiaoli and Zhu Xuezhong constructed a patent evaluation index system from the perspectives of technology, market and rights. They included 17 evaluation indicators and used the hierarchical method and fuzzy comprehensive theory to establish an evaluation model. In 2012, the Patent Value Analysis Index System Operation Manual, jointly published by the State Intellectual Property Office and the China Technology Exchange, established a patent value analysis index from the perspective of patent legal value, technical value and economic value, and contained 18 evaluation indicators. However, the indicators of the impact value and the importance of the above patent value evaluation system are mainly determined by the experts' assumptions about the importance of the indicators but not tested. The subjective factors have a greater impact. Therefore, it is necessary to establish a more targeted high-value patent evaluation system.

Summary of the invention

In view of this, the present disclosure provides a patent value evaluation method, a method for constructing a patent value evaluation model, and a computer readable storage medium, which can generate an objective and accurate evaluation model in a targeted manner, and improve the evaluation of patent value. accuracy.

According to a first aspect of the present disclosure, a method for constructing a patent value evaluation model is provided, which comprises: selecting parameters affecting patent value; selecting patent data samples according to a preset patent value evaluation standard; The selected patent data samples establish an evaluation formula that reflects the relationship between the selected parameters and the patent value.

Preferably, the parameters include at least one of: parameters related to the type, status, content, review process of the patent; parameters related to the citation and citation of the patent; parameters related to the layout of the patent; Parameters related to the maintenance of patents; parameters related to the inventor and applicant of the patent; parameters related to the operation of the patent; and parameters related to patent disputes and disputes.

Preferably, the selecting parameters affecting the patent value includes: selecting a factor affecting the patent value and a sub-parameter for evaluating the factor; evaluating the factor according to the sub-parameter, and using the evaluation result as the selected one. parameter.

Preferably, the method further comprises: after selecting the patent sample, establishing a relationship map between the selected parameter and the patent value based on the selected patent data sample; and the evaluation formula is established based on the relationship map of.

Preferably, the selecting the patent data sample according to the preset patent value evaluation standard comprises: selecting a patent that meets the preset first condition as the first group of patent data samples; and selecting a patent that meets the preset second preset condition As a second set of patent data samples.

Preferably, the second set of patent data samples are selected from a first set of patent data samples, the first condition including all inventions and utility models described in the national patent database, the second condition including the following At least one of: a patent infringement lawsuit and a successful patent; a patent that maintains the payment until the expiration of the patent period; a patent that is put into operation and receives a return; a Chinese patent that is requested to be invalid and sentenced to remain valid; and an industry-standardized patent; A patent that has been used more than a preset number of times; a patent applied for in multiple countries.

Preferably, the evaluating formula for establishing a relationship between the selected parameter and the patent value based on the selected patent data sample comprises: calculating a distribution of the selected parameter in the first set of patent data samples; calculating the selected The distribution of parameters in the second set of patent data samples; the comparison between the distribution of the selected parameters in the first set of patent data samples and the distribution of the selected parameters in the second set of patent data samples; The comparison result fits the relationship between the selected parameter and the comparison result as an evaluation formula that reflects the relationship between the selected parameter and the patent value.

Preferably, the evaluation formula for establishing a relationship between the selected parameter and the patent value based on the selected patent data sample comprises: extracting a patent corresponding to each value of the parameter; performing a patent corresponding to each value The evaluation of patent value, and based on the evaluation result, the influence of the value of the parameter on the patent value; the evaluation formula of the relationship between the parameter and the patent value based on the influence of each value of the parameter on the patent value .

Preferably, the patent data sample comprises a patent data sample of a single patent or a patent data sample of a composite patent converted from a plurality of patents.

Preferably, there are multiple selected parameters, and the evaluation formula for establishing a relationship between the selected parameter and the patent value based on the selected patent data sample includes: selecting, for each selected parameter, respectively The patent data sample establishes a sub-evaluation formula that reflects the relationship between the parameter and the patent value; the weighted sum of the sub-evaluation formula is used as an evaluation formula for the selected all parameters.

Preferably, the selected parameters are divided into multiple groups, and the evaluation formula for establishing a relationship between the selected parameters and the patent value based on the selected patent data samples includes: for each selected group of parameters, Based on the selected patent data samples, a sub-evaluation formula reflecting the relationship between the set of parameters and the patent value is established; the weighted sum of the sub-evaluation formulas is used as an evaluation formula for all selected parameters.

Preferably, the method further comprises: receiving other parameters; establishing at least one other model for evaluating the patent value based on the other parameters, combining the other models with the established evaluation formula to obtain a final constructed patent value evaluation model.

According to a second aspect of the present disclosure, a computer readable storage medium is provided for storing instructions that, when executed by a processor, cause a processor to perform a build patent value evaluation model according to any of the above Methods.

According to a third aspect of the present disclosure, a patent value evaluation method is provided, which comprises: constructing a patent value evaluation model; evaluating a target patent according to the constructed patent value evaluation model; The patent evaluation result is used to optimize the patent value evaluation model, wherein the construction patent value evaluation model includes: selecting parameters affecting the patent value; selecting patent data samples according to the preset patent value evaluation criteria; The patent data sample establishes an evaluation formula that reflects the relationship between the selected parameters and the patent value.

Preferably, the constructed patent value evaluation model includes a first model and a second model different from the first model, and the patent value evaluation model is optimized according to the evaluation result of the target patent, including: the first model and And/or the second model is evaluated as a target patent whose value is in the desired range as a reference patent; and the reference patent is used to update the patent data sample for constructing the first model and/or the second model.

Preferably, the constructed patent value evaluation model includes a first model, and the optimization of the patent value evaluation model according to the evaluation result of the target patent includes: the first model is evaluated as a target patent whose value is within a desired range. Reference patents; and using the reference patents to update patent data samples used to construct the first model.

Preferably, the constructed patent value evaluation model includes a first model and a second model different from the first model, and the patent value evaluation method further includes: pre-selecting the first model and the second model respectively The verification data is evaluated to obtain a first evaluation result and a second evaluation result; the first evaluation result is compared with the second evaluation result; and the first model and/or the second model are adjusted based on the comparison result.

Preferably, the constructed patent value evaluation model includes a first model, and the patent value evaluation method further includes: using the first model to evaluate the pre-selected verification data to obtain the first evaluation result; The patent data sample used by the model obtains a second evaluation result; the first evaluation result is compared with the second evaluation result; and the first model is adjusted based on the comparison result.

Preferably, the patent value evaluation method further comprises: after evaluating the target patent according to the constructed patent value evaluation model, normalizing the evaluation result to be converted into an evaluation score value within a preset range. .

Preferably, the target patent includes a patent group, and the evaluation of the target patent according to the constructed patent value evaluation model includes: evaluating each patent in the patent group according to the constructed patent value evaluation model; The evaluation results of each patent in the group were mathematically calculated to obtain evaluation results for the patent group.

According to a fourth aspect of the present disclosure, there is also provided a computer readable storage medium storing instructions that, when executed by a processor, cause a processor to perform the patent value evaluation method according to any of the above .

DRAWINGS

In order to more clearly illustrate the technical solutions of the embodiments of the present disclosure, the drawings of the embodiments will be briefly described below. It is obvious that the drawings in the following description relate only to some embodiments of the present disclosure, and are not to limit the disclosure. .

FIG. 1 shows a schematic flow chart of a method of constructing a patent value evaluation model in accordance with an embodiment of the present disclosure.

2A is a diagram showing the probability that a number of patents of different claims in the Chinese patent appear in the first patent data sample according to an embodiment of the present disclosure.

2B is a diagram showing the probability that a number of patents of different claims in the Chinese patent appear in the second patent data sample according to an embodiment of the present disclosure.

2C shows a comparison of the probability of occurrence of patents of different claims in the Chinese patent in the first and second patent data samples in accordance with an embodiment of the present disclosure.

3A illustrates the probability of occurrence of a number of patents of different claims in a first patent data sample in accordance with an embodiment of the present disclosure.

3B illustrates the probability of occurrence of a number of patents of different claims in a U.S. patent in a second patent data sample in accordance with an embodiment of the present disclosure.

3C illustrates a comparison of the probabilities of the number of patents of different claims in the U.S. patents in the first and second patent data samples in accordance with an embodiment of the present disclosure.

Figure 4 shows the fitting curve of the relationship between the number distribution of Chinese high-value patents and ordinary patent claims.

FIG. 5 shows a schematic flow chart of a patent value evaluation method according to an embodiment of the present disclosure.

FIG. 6 shows a schematic diagram of a comprehensive evaluation result of patent value according to an embodiment of the present disclosure.

FIG. 7A is a graph showing the value analysis of the patents of the first to the eighteenth China Patent Awards by using the patent value evaluation method of the embodiment of the present disclosure.

FIG. 7B is a graph showing the value analysis of the ETSI Chinese standard essential patent by the patent value evaluation method of the embodiment of the present disclosure.

FIG. 7C illustrates the evaluation of the comprehensive value of a single patent and the evaluation of the patent price of each dimension by the patent value evaluation method according to an embodiment of the present disclosure.

FIG. 8A shows a schematic diagram of a user interface for patent value evaluation using a patent value evaluation method according to an embodiment of the present disclosure.

FIG. 8B shows a schematic diagram of the applicant's patent layout using the patented value evaluation method according to an embodiment of the present disclosure.

detailed description

The technical solutions of the embodiments of the present disclosure will be clearly and completely described in the following description of the embodiments of the present disclosure. It is apparent that the described embodiments are part of the embodiments of the present disclosure, and not all of the embodiments. All other embodiments obtained by a person of ordinary skill in the art based on the described embodiments of the present disclosure without departing from the scope of the invention are within the scope of the disclosure.

The present disclosure provides a method for constructing a patent value evaluation model, a patent value evaluation method, and a computer readable storage medium. Instead of using an expert scoring method, the method of using big data analysis separately obtains the patent value of each parameter. Influence, and thus establish an evaluation formula of patent value, and obtain a patent value model. On the one hand, it can generate objective and accurate evaluation models in a targeted manner, and on the other hand, it can improve the accuracy of patent value evaluation.

FIG. 1 shows a schematic flow chart of a method of constructing a patent value evaluation model in accordance with an embodiment of the present disclosure.

In step S101, parameters affecting the patent value are selected.

An example of the parameters affecting the patent value is given in Table 1 below. The desired parameters can be selected from Table 1 as needed. In order to ensure the comprehensiveness of the model establishment, at the beginning of the establishment of the patent value evaluation model, all kinds of factors that may affect the patent value should be found out, including the evaluation indicators commonly used internationally, and also according to the Chinese patent protection. The indicators summed up by the characteristics of the situation include both direct parameters directly related to the patent itself, as well as indirect parameters related to the technical fields of the applicant, the agency and the patent.

Table 1

In step S102, a patent data sample is selected according to a preset patent value evaluation standard.

In this step, the patent that meets the preset first condition may be selected as the first group of patent data samples, and the patent that meets the preset second preset condition is selected as the second group of patent data samples. For example, samples of known high-value patents and samples of common patents may be selected according to the general criteria for evaluation of patent value.

For example, a patent that meets at least one of the following criteria may be selected as a patent sample of known high value: 1. a patent that has participated in an infringement lawsuit and obtained a substantial victory; 2. a longevity patent that maintains the payment until the expiration of the patent period; 3. the patentee Putting patents into operation and obtaining high-value patents through implementation, licensing, or transfer; 4. Patentes that are submitted for invalidation and are judged to remain valid; 5. Be incorporated into standards or form factual standards through practical application to become industry norms Basic patents; 6. patents that have been used more than a predetermined number of times; 7. patents filed in multiple countries, such as patents filed in multiple countries through the PCT, Paris Convention or other means; 8. Other technologies for the industry or Patents that have a significant impact on the market.

On the one hand, because the level of development of patent litigation in China is far from that in the United States, it is impossible to meet the analytical needs in terms of the number and quality of data samples that can be selected. Taking the maintenance period of patents as an example, the number of Chinese invention patents that have been paid until the expiration of the patent period has reached 2,156 as of December 2017, 20% of which belong to Sinopec Group. It can be seen that the patents expired in China's current patent period mainly come from state-owned enterprises. Less relevant to specific values. On the other hand, the lack of relevant data can also cause obstacles in the extraction of high-value patent data samples. For example, patent holders can clarify the patent application and operational value of patents in China, and there is not enough effective data support. Therefore, when we choose China's high-value patent data samples, we should proceed from reality and cannot copy the theory.

As an illustration, the present embodiment uses 2,305 invention patents selected in the Chinese patent database that are invalid and maintained effective as a simplified sample of high-value patent analysis. Relatively speaking, the Chinese patent is invalidated and the number of patents that have been validated is sufficient. The cost of human and financial resources is invalid. It indicates that the patent has been applied in the industry or has an impact on the industry. value.

As a common patent sample for comparison, all inventions and utility model patents of the Chinese patent database can be directly selected, and the statistical data fully reflects the distribution of the number of claims in the whole library patent.

Certainly, the embodiment of the present disclosure is not limited thereto, and the selection criteria of the patent data sample may be adjusted according to requirements. For example, all patents of an applicant may be selected as a sample of common value patent data, and all patents of the applicant are selected. A patent that meets the pre-determined conditions is used as a sample of high or low value patent data. In some embodiments, the selection of high or low value patent data samples may also be selected from a range of generic value patent data samples, or may be selected from other ranges.

In addition, the above patent data sample may be a data sample of a common single-piece patent, or may be processed into a single-piece patent form of a composite patent data sample by a plurality of patents. For example, a family of patents typically includes multiple pieces of patents that are related to each other, such as patents that are filed in multiple countries based on the same priority. In some cases, the feature parameters of each patent in the family patent can be integrated in the database to be converted into a single patent form for retrieval and analysis. For example, an applicant for a single patent converted from a family patent may be a collection of individual applicants of the same family patent, and the number of claims may be the sum or average of the number of claims of each family patent. The patent data sample of this embodiment covers this situation. For example, the patent data sample of the same family patent converted into a single patent form can be extracted to establish a patent value evaluation model to evaluate the value of the family patent. In this case, because of the difference in the relationship between the single patent and the patent family in terms of parameters, weights, and various parameters, this embodiment can be used to evaluate the patent family data of the global family for the value evaluation model of the same family patent. Based on the analysis, it is not a simple accumulation of the individual patent value scores in the family or simply accumulating the corresponding parameters of all patents in the patent family and then evaluating them according to the value evaluation model of a single patent.

In step S103, a relationship map between the selected parameter and the patent value is established.

After the analysis sample of the high-value patent is determined, the association relationship with the patent value can be calculated one by one according to the previously selected evaluation parameter list. In this embodiment, the number of claims is selected as a parameter affecting the patent value, and this is taken as an example for explanation.

According to the provisions of the Patent Law, the scope of protection of patents for inventions or utility models is subject to the contents of the claims. In fact, the scope of each claim is described by words and is difficult to quantify. Although the number of claims does not directly reflect the scope of patent claims, the number of claims may reflect the scope of protection of patent claims and the stability of rights to some extent. On the one hand, the number of claims is large, which may indicate that the technical solution of the patent is more complicated, or the invention content is more abundant. The greater the number of independent claims, the greater the perspective of protecting the same inventive concept, or the more technical solutions derived from the inventive concept, the greater the scope of patent protection; the number of dependent claims The more the number of technical solutions derived from the corresponding independent claims, or the more detailed the improvement of the technical solutions, the greater the scope of protection can be retained as far as the independent claims cannot be maintained, and the rights should be maintained as much as possible. Stability. On the other hand, as mentioned above, according to the provisions of the Chinese Patent Law, when the number of claims exceeds 10, the surcharge is required at the time of application, and at the same time, the agent pays more labor in the writing process, and the patent agent pays. Fees may also increase, resulting in a relatively high cost to the applicant when applying for a patent. Therefore, theoretically, patents with a large number of claims will have a higher probability of forming high-value patents.

In order to verify the above theory, the following is an example of a Chinese patent and a US patent.

In this step, a comparison table between the high value patent and the ordinary patent can be established for the quantity distribution of the claims. Calculate high-value patent samples and whole-library patents separately (Note: The total library patents used in this example include all invention applications and all utility model patents in China as of October 25, 2017). The number of claims is equal to n (n is The number of patents with a natural number greater than or equal to 1, and the proportion of the number of patents in high-value patents and ordinary patents, respectively.

2A is a diagram showing the probability of occurrence of patents of different claims in the Chinese patent in the first patent data sample (general patent) according to an embodiment of the present disclosure. 2B is a diagram showing the probability of occurrence of patents of different claims in the Chinese patent in the second patent data sample (high value patent sample) according to an embodiment of the present disclosure. 2C shows a comparison of the probability of occurrence of patents of different claims in the Chinese patent in the first and second patent data samples in accordance with an embodiment of the present disclosure.

As can be seen from Figures 2A and 2B, for both the general patent and the high value patent, when the number of claims increases from 1, the icons all show a tendency to rise first and then decrease. Only ordinary patents begin to decline when the number of claims is greater than 3, while high-value patents begin to decline after the number of claims is greater than 10.

Next, the probability of occurrence of high-value patents with a number n of patent claims in high-value patents and the probability multiples of ordinary patents with a number of patent claims n in the whole library are calculated, and the relationship between the two is obtained. As shown in Figure 2C. As can be seen from FIG. 2C, when the number of claims is less than or equal to 6, the number of claims of high-value patents is less than the number of common patents. When the number of claims is greater than 6, the proportion of high-value patents is generally higher than ordinary. The value, and the multiples that are higher, tend to increase linearly. Compared with ordinary patents, the number of claims for high-value patents is relatively large, and the greater the number of claims, the greater the probability that high-value patents will occur, which is consistent with our original expected trend.

To more clearly show the relationship between the patent value and the number of claims, we can simply fit the above scatter plot into a one-dimensional linear curve: y=0.085x+1.0695, and its slope of 0.085 represents the number of claims. The degree of relevance of patent value.

In fact, the relationship between the value of patents and the number of claims does not always show a linear growth trend. As can be seen from the above figure, when the number of claims is small (if the number of claims is less than 10), the trend of patent value increases with the number of claims is steep, and as the number of claims continues to increase, the number is The impact of patent value will gradually decrease and tend to be flat. Therefore, when constructing the relationship between patent parameters and patent value, we will give a more complex and accurate fitting strategy.

As can be seen from Figures 2A through 2C, when the number of claims rises to a certain number (in this example, when n is greater than 23), the distribution of data presents a discrete trend. This is because the above-mentioned high-value patent samples are only used as case demonstrations, so the number of samples is small, and when the number of claims is large, there is a case where the number of samples is missing. If the data sample of a high-value patent is large enough, the above-mentioned discrete phenomenon will gradually weaken, but it will not disappear, because in fact, high-value patents also have fewer claims, and low-value patents also have a larger number of claims. The number of claims can only reflect the probability of a high-value patent occurring, and cannot absolutely represent the level of patent value.

In addition, as can be seen from FIGS. 2A to 2C, when the number of claims is 10, the proportion of the distribution of high-value patents is "diving" compared with the ordinary patent, according to the trend of the data before and after, when the number of claims is 10 The proportion of high-value patents should be about 2 times that of ordinary patents, while the statistical chart is only 0.75, which is smaller than the claims number of 6, 7, 8, and 9, and less than the number of claims. 10% of high-value patents. The value of a patent that is not a claim number of 10 is necessarily small in probability, and the problem still arises from the patent sample. According to the screening conditions of the high-value patents described above, the vast majority of the samples obtained are authorized invention patents. Invention patents need to undergo substantive examination before authorization. In the substantive examination, most patent claims will be partially modified by the examiner's question, resulting in a difference in the number of patents for the final authorization from the original application (usually Is reduced). The “full library” patent representing the common patent sample includes all the invention applications and utility model patents in China, and these patents have not been substantively examined. In view of the provisions of the Chinese patent charging standard for the surcharge of claims, many applicants or agents are accustomed to limit the number of claims to 10 or 10, thus resulting in the number of claims in unlicensed patents. A patent with a probability of 10 has a higher distribution probability, making the statistics look abnormal. Similarly, in the above scatter plot, the statistical values of the number of claims of 20, 30, 40, etc., which are 0, are more or less anomalous, and all have some relationship with this or some (some people I like to "make integers" when writing claims. When we actually construct the patent value evaluation formula, we can more realistically restore the relationship between patent parameters and patent value by ignoring these abnormal data points.

Of course, the function reflected by a simple high-value patent sample model is not necessarily very precise. The solutions are as follows: one is to increase the number of high-value patent samples and the purity of the high-value patents contained therein; Construct different high-value patent sample models, perform back-to-back calculations separately, and then compare them separately to find out the causes and adjust accordingly. Third, use the preliminary constructed high-value patent model to group multiple sets of randomly sampled patent samples. The accuracy of the grouping results is analyzed, and the weights of related parameters are manually adjusted. In short, the construction of a sample model of high-value patents is a very complex process that requires repeated calculations, comparisons, adjustments, and verifications.

In addition, considering that the patent database as a whole is a dynamic and ever-increasing data set, the results of the big data statistics and the laws of reflection will continue to change over time. Therefore, when establishing the patent value evaluation model It is also necessary to periodically check and adjust the model.

In the process of establishing the relationship between the number of patents and the evaluation parameters, we did not make special settings based on the parameters of Chinese patents by expert scoring. However, this relationship map truly reflects the true characteristics of Chinese patents. The basis for the establishment of the map is China's patent data and high-value patents extracted from Chinese patent data, so it will not be affected by the characteristics of patents in other countries and regions.

In the same way, we can create a relationship map between the number of US patents and evaluation parameters, and by comparison, find the difference in the relationship between the number of patents and the value of patents in these two different countries. This will be explained below with reference to Figs. 3A to 3C. As an example, we select all patents filed and published in the US since 2000, as a basic sample, and select the standard essential patents (a total of 18,393) related to the ETSI (European Telecommunications Standards Institute) standard as high value. The patent sample, after analysis using the above method, obtains the following relationship map:

3A illustrates the probability of occurrence of a number of patents of different claims in a first patent data sample in accordance with an embodiment of the present disclosure. 3B illustrates the probability of occurrence of a number of patents of different claims in a U.S. patent in a second patent data sample in accordance with an embodiment of the present disclosure. 3C illustrates a comparison of the probabilities of the number of patents of different claims in the U.S. patents in the first and second patent data samples in accordance with an embodiment of the present disclosure.

As can be seen from Figure 3C, the probability relationship between the distribution of high-value patents and common patent claims is:

y=0.0297x+0.5452 (1)

From equation (1), it can be concluded that the number of claims has a relevance to patent value of 0.0297.

Comparing Figure 2 with Figure 3 can be seen:

(1) Comparing the probability that the patents of China and the United States with the claim n appear in all samples, it can be seen that the majority of Chinese patents are distributed in the interval of the number of claims less than or equal to 10, and the number of claims is 1 and 2 respectively. The proportion of patents reached 9.6% and 9.8% respectively. Except for the claim equal to 10, the most probable occurrence is the patent with the number of claims equal to 3. The number of patents is gradually decreasing in the range of the number of claims equal to 3 to 9, but when the number of claims is greater than 10, the number of patents A cliff-like sudden drop occurs, and when the number of claims is greater than 12, the proportion remains substantially below 1%. In the case of US patents, when the number of claims is less than 10, the number of patents shows a slower upward trend. In the interval where the claims are equal to 10 to 20, the line is basically flat. When the number of claims is greater than 20, the number of patents is slow. The proportion of patents in the range of 20 to 30 is basically greater than 1%. It shows that the number of claims for common patents in the United States is mostly kept in a large range of values, and the overall level of patent writing is higher than that of Chinese patents. In addition, when the number of claims is equal to 20, the distribution of high-value patents and ordinary patents in the United States has a clear sudden rise, which is obviously directly related to the collection rules of the US patent claims surcharge.

(2) Comparing the distribution probabilities of high-value patents with Chinese and American claims n, we can see that the number of Chinese high-value patent claims is basically increasing within 10, when the number of claims is greater than 10, although There is a significant drop in the quantity distribution, but it will slowly decrease from around 3%. When the number of claims is equal to 21, it will remain above 1%. With the increase in the number of claims, the number of high-value patents in the United States still basically rises and then declines. However, in general, the proportion of patents with a large number of claims is higher than that of high-value patents in China. It shows that although China's high-value patents have a larger number of average claims than Chinese ordinary patents, there is still a certain gap compared with US high-value patents.

(3) Generally speaking, with the increase in the number of claims, the average proportion of high-value patents in the United States is also increasing, but the growth rate (slope) is significantly lower than that of Chinese patents, especially when the number of claims is less than 20, The number of high-value patents and ordinary patents has no obvious advantage in probability distribution, indicating that for Chinese patents, the number of claims is more obvious in terms of patent value, or that Chinese high-value patents are in rights relative to ordinary patents. The difference in the number of requirements is more significant.

In step S104, an evaluation formula for constructing a relationship between patent parameters and patent value is constructed.

Although in theory, the more relevant parameters are included in the model, the more detailed and detailed the patent value will be, the more accurate the evaluation results should be. However, it should be considered that the increase of evaluation indicators will increase the complexity and calculation of the model, and the relationship between some indicators (such as patent examination duration) and patent value is very complicated. It needs to be analyzed specifically for specific situations, only mechanically. The results of applying systematic data statistics sometimes have the opposite effect on the judgment of patent value; and some indicators that have an impact on research problems but are temporarily difficult to fully obtain (such as the importance of invalid claims in partial invalid review decisions) will also Reduce the feasibility of this research method. These patents need to be selectively discarded on the basis of careful analysis of the experimental results. Therefore, we need to consider the above factors and the actual map test results to select the evaluation parameters used in the model.

After determining the selected evaluation parameters, the next step is to analyze the actual characteristics of each parameter and construct an evaluation formula for the relationship between the parameter and the patent value. The following is still taking the claim quantity parameter as an example to illustrate the construction process of the number of claims and the evaluation formula of the patent value.

(1) Construction of patent value evaluation formula for direct parameters satisfying curve relationship

Figure 4 shows the fitting curve of the relationship between the number distribution of Chinese high-value patents and ordinary patent claims. By performing an in-depth analysis of the scatter plot of Figure 4, the following nonlinear mathematical model can be fitted:

Among them, Vi refers to the contribution of the i-th patent parameter to the patent value. M is the maximum value that Vi can take, p is the parameter value, and K is the value of p when Vi tends to be gentle.

In the fitting curve of the relationship between the number distribution of Chinese high-value patents and ordinary patent claims shown in FIG. 4, p can represent the number of current patent claims, and M represents the maximum value of the influence of the claim parameters on the patent value, that is, when When the number of claims p is greater than the threshold K, the increase in the number of claims is no longer reflected in the value of the patent. In this embodiment, the value of K is 50.

Compared with the linear fitting result of the above equation (1), equation (2) can more accurately reflect the influence of the number parameter of the Chinese patent claim on the patent value.

Of course, the above fitting method can also be used for other parameters such as the number of patent citations and the number of patents of the same family. It should be noted that the shape of the influence curve on the patent value exhibited by different parameters and the reasons for the influence are different, and need to be separately performed. Targeted analysis and design, in some cases even need to be segmented to give the corresponding functional relationship.

(2) Construction of patent value evaluation formula for direct parameters that do not satisfy the curve relationship

Not all parameters that may affect the value of a patent can be constructed by fitting the above curves. For example, the type of applicant (enterprise, college, other institution, individual), the type of patent (invention, utility model), validity (effective, trial, invalid) or legal status (litigation, licensing, transfer, invalid, etc.) ) will have an impact on the value of the patent, but these parameters are difficult to analyze through the statistics of the variables. For this type of parameter, the distribution probability of the parameters in high-value patents and general patents can also be statistically analyzed by means of big data analysis, and the influence of each case on the value of the value is estimated and the corresponding influence weight is given. For example, taking the applicant type as an example, the patent corresponding to each value of the parameter can be extracted, and the patent value evaluation of the patent corresponding to each value can be performed (here, the established model can be used for evaluation, for example, according to the above satisfaction curve) A preliminary model of other parameters of the relationship is constructed to evaluate, or otherwise evaluated), and based on the evaluation result, the weight of the value of the parameter is added to the patent value, based on the influence of each value of the parameter on the patent value The weights are used to obtain an evaluation formula that reflects the relationship between parameters and patent value.

Taking the applicant type parameter as an example, the applicant can be divided into groups according to the applicant, the enterprise, the college, other institutions, and individuals, and the preliminary model constructed by parameters such as the number of claims, the number of patent citations, and the number of patents of the same family can be used. Obtaining the value of each patent in each group, and obtaining the comprehensive value of the patent of the group, the manner of obtaining the comprehensive value may refer to the following content described in step S105. The influence weight of the corresponding parameter value can be designed according to the comprehensive value of each group, and the influencing factors of the parameter are included in the patent value evaluation model.

Taking the applicant type parameter as an example, it is assumed that the influence of the four types of applicants, such as enterprises, universities, other institutions and individuals, on the patent value is a, b, c, and d, respectively. , in the evaluation of patent value, according to the following rules:

If applicant type = business: V1 = a

If applicant type = college: V2 = b

If applicant type = other institutions: V3 = c

If applicant type = individual: V4 = d

Among them, V1, V2, V3 and V4 respectively represent the contribution of the four applicant types of enterprises, universities, other institutions and individuals to the patent value.

(3) Construction of patent value evaluation formula for indirect parameters

In addition to the direct parameters, the above Table 1 also lists 14 indirect parameters such as the “Patent Agency Comprehensive Patent Application Quality Index”. Each indirect parameter is obtained by comprehensively calculating a number of sub-parameters affecting its value. In some embodiments, each indirect parameter is also obtained by a different value evaluation model. For example, the indirect parameter of the “Patent Agency Comprehensive Patent Application Quality Index” is the score of the patent comprehensive value scored by the agency, the number of invention patents represented by the agency and its proportion, the patent grant rate of the invention, and the average of the patent. A plurality of parameters such as the number of claims are comprehensively evaluated. When constructing the evaluation formula for indirect parameters, the factors affecting the patent value can be selected. At the same time, the sub-parameters for evaluating the factors can be selected, and the factors are evaluated according to the sub-parameters, and the evaluation results are taken as the selected indirect parameters. Specifically, indirect parameters can be obtained by the following equation:

Where p _i represents the i-th indirect parameter, s represents the number of sub-parameters affecting the number of indirect parameters, r _t refers to the contribution of the t-th sub-parameter to the indirect parameter value, and w _t represents the weight of the t-th sub-parameter .

The relationship between indirect parameters and patent value can be expressed as:

V _i =f(p _i ) (4)

Where V _i represents the contribution of the i-th indirect parameter to the patent value, and f() represents the functional relationship.

It should be pointed out that the indirect parameters listed in the attached table of this chapter also include the “Preliminary Value Evaluation Index of a Single Patent”. The method for obtaining this parameter is as follows: 1. Initial use of other indicators except this indicator to construct the patent value. The evaluation model obtains the preliminary value evaluation index of the patent; 2. Iteratively optimizes some direct parameters or indirect parameters for the preliminary value evaluation index, and obtains the optimization model of patent value evaluation; 3. Use the optimization model to obtain the patent final Patent value.

In step S105, a comprehensive evaluation formula of patent value is established.

After the above evaluation formulas of direct parameters and indirect parameters are established, a comprehensive evaluation formula for patent value can be constructed.

The current parameter system for evaluating patent value is roughly divided into several structural types such as one-dimensional structure, linear structure and tower structure. The one-dimensional structure is evaluated by a single parameter, which is obviously not suitable for evaluating the value of the patent. Linear structures use multiple parameters, but these parameters are in a side-by-side relationship with each other. Because there are many factors affecting the judgment of patent value, the linear structure will make the system too complicated, and it is difficult to grasp the relationship between the parameters, and it may cause the partiality of the situation to reduce the reliability of the analysis results. According to the characteristics of the parameters used to evaluate the patent value, the tower evaluation structure can be selected, that is, all the evaluation parameters are divided into multiple levels according to the angle of the patent value. For example, parameters can be divided into groups, each group corresponding to one evaluation dimension. In this embodiment, the parameters can be divided into three groups according to three dimensions: technology (technical advancement), law (technical stability), and market (protection scope). The technical stability mainly considers whether the patent has undergone substantive review and Obtain authorization or maintain effective, whether it has been tested by market infringement litigation, pledge and other market operation activities, and whether there are overseas-licensed family patents; technological advancement mainly considers the citation of patents and their family patents worldwide, patent investment The number of R&D personnel, the technical field involved in patent technology (the number of IPC classifications), the transfer or licensing of patents, etc.; the scope of protection mainly refers to the number of patent claims, the remaining protection period, the number of countries or regions where the patent is laid out, and so on. Since the selected parameters are hierarchical, the weights of the items in each level are controllable, and can be expanded correspondingly as the parameters increase, so it is more flexible and less restrictive.

The comprehensive value of a patent is obtained by the sum of the combined values of the above three dimensions:

Wherein, when i ranges from [1, n1], [n1+1, n2], [n2+1, n], respectively, V _i represents technology (technical advancement), law (technical stability), The contribution of patent parameters in the three dimensions of the market (protection scope) to the patent value. The value of each dimension can be controlled within the desired range by means of extreme value definition.

In addition, based on the above-mentioned Chinese patent value evaluation model, a value evaluation model for a patent collection (such as all Chinese patents applied for by an applicant) can be established, thereby integrating the multiple different sets. The value of patents is evaluated and compared.

In addition, in addition to the parameters listed in Table 1, user-defined parameters can also be set. In this case, the user can select other parameter types as desired, and if necessary, provide a corresponding data list as a patent data sample. portion. When constructing the patent value evaluation model, some or all of the parameters can be selected from Table 1 to construct the preliminary patent value model, and another patent value evaluation model based on user-defined parameters can be constructed to make it and the preliminary patent. The value model is combined to obtain the final constructed patent value evaluation model.

In addition, the value of the patent is reflected in the strategic value, technical value, legal value, market value and economic value of the patent. The patent value evaluation model of the above embodiment can give the value evaluation result at least part of the level. For the level that cannot be covered by the above embodiment, the patent value evaluation model can be given by other means or even the evaluation result can be directly obtained, and the final patent value evaluation is obtained by combining with the patent value evaluation model of the above embodiment. For example, based on the evaluation results of the above-mentioned patent value evaluation model on technical value, legal value and market value, the evaluation of the relevant evaluation dimension of strategic value or economic value can be increased, thereby expanding the application scope from macro analysis to case-specific. The value of the analysis. The strategic value and economic value of a patent depend to a large extent on some special attributes of the patent itself, such as the implementation and application of patent-related technologies or products, the circumvention of patent technology, confidentiality or substitutability, evidence of patent infringement. The difficulty of obtaining, as well as the cost of asset replacement for the applicant and the patent, the recent transaction price of the same or similar assets on the market, the excess return from the patent assets, etc., and many of these attributes are personalized, often It is difficult to obtain through traditional patent information or legal information, and it is necessary to assign values separately for specific situations.

It should be noted that the above embodiment is only for explaining the principle of constructing the patent value evaluation model, and more complicated influencing factors should be considered in the actual construction process. For example, there may be strong correlations between some parameters (such as the patent authorization and the authorization of the same family patent), which may cause repeated weighting. In addition, the model in this embodiment is a comprehensive evaluation model for invention and utility model patents, wherein the parameters of the invention application and the utility model that have not been actually examined and the invention patents after the trial are different, the same parameter The value of the patents reflected is not the same. For these complicated situations, multi-angle judgments can be made on the logical relationship, and selective merger or appropriate adjustment of the weights can be made according to different situations. Therefore, in the actual operation, the relevant indicators can be traded off and integrated in combination with the results of the correlation test.

FIG. 5 shows a schematic flow chart of a patent value evaluation method according to an embodiment of the present disclosure.

In step S201, a patent value evaluation model is constructed. In this step, the patent value evaluation model can be constructed by the method described above with reference to FIGS. 1 to 3. The number of models built can be one or more. For example, a first model can be constructed based on a set of patent data samples, and a second model can be constructed based on another set of patent data samples for use in subsequent steps.

In step S202, the patent value evaluation model is optimized. Ways to optimize include but are not limited to iterations and comparisons.

For example, assuming that a model is constructed in step S201, that is, the first model, the target patent evaluated as the value in the desired range by the first model can be used as a reference patent, and the reference patent is used to update the first model. Patent data sample. The desired range here may be selected as needed, for example, above a preset value, below a preset value, and/or within a preset range. For models used to screen high-value patents, patents above the default value can be used as reference patents to update patent data samples; for models used to screen low-value patents, patents below the default value can be used as reference patents. To update the patent data sample. In some embodiments, also in the case of constructing a model, the model can also be optimized by using the model to evaluate pre-selected verification data to obtain a first evaluation result, and to evaluate the construction of the model. The patent data sample used is used to obtain a second evaluation result, the first evaluation result is compared with the second evaluation result, and the model is adjusted based on the comparison result. Here, the operation of evaluating the patent data sample to obtain the second evaluation result may be performed according to experience or a preset standard, or may be performed by other means or by using other tools. After the comparison, the model is optimally adjusted according to the deviation between the two evaluation results, thereby improving the accuracy.

Similarly, assuming that two models, namely the first model and the second model, are constructed in step S201, the target patent evaluated by the first model and/or the second model as the value range can be used as a reference patent. The reference patent updates a patent data sample for constructing a first model and/or a second model. Similarly, the so-called desired range can also be selected as needed, and will not be described again here. As another example, a set of patent data for verifying the model may also be prepared, and the first model and the second model are respectively used to evaluate the patent data and obtain the evaluation result, and the evaluation result of the first model and the evaluation of the second model are obtained. The results are compared to find out the cause of the difference and the first model and/or the second model are adjusted accordingly.

In some embodiments, when the target patent is evaluated by using the model, if the target patent meets the preset condition, the weight of the corresponding sub-evaluation formula in the model used for the target patent is increased or decreased. . For example, for high-value patents, it is clear that the longer the remaining validity period, the higher the patent value. However, for low-value patents, the value of the patent has no obvious correlation with the length of the remaining validity period. If the weight is applied in the same way as the high-value patent, it will inevitably affect the objectivity of the evaluation result. Therefore, through the above iterative method, after the target patent is evaluated as a high-value patent, it is considered necessary to strengthen the contribution of the remaining life parameter to the patent value, and increase the weight of the corresponding sub-formula, and vice versa. Then, use the optimized model to re-evaluate the target patent, and get a more accurate value evaluation result.

In step S203, the value of the target patent is evaluated using the constructed patent value evaluation model. For example, if the patent value evaluation model obtained in step S201 is constructed based on three sets of parameters, the three sets of parameters respectively correspond to the evaluation criteria of the three levels of technical stability, technological advancement, and protection scope. Then, in this step, the model can be used to obtain the evaluation results of the target patents at the three levels, and the evaluation results are weighted and summed to obtain the comprehensive evaluation result, according to which the search results can be sorted according to the patent value.

In this step, the appropriate model can be used to evaluate the value of the target patent as needed. The target patents here include, but are not limited to, a single patent or patent group.

For example, if the target patent is an ordinary single-piece patent, the value of the patent value evaluation model constructed by the data sample of the single patent can be used to evaluate the value. If the target patent is a composite patent in the form of a single piece processed by the same family patent, the value of the patent value evaluation model constructed with the corresponding composite patent data sample can be used to evaluate the value. If the target patent is a patent group (such as a patent group divided by the applicant, IPC classification number, applicant territory, family patent, etc.), each patent in the group can be individually evaluated, and the evaluation result can be mathematically calculated. The results of the evaluation for the patent group are obtained, including but not limited to averaging, geometric averaging, square averaging (root mean square mean, rms), harmonic mean, weighted average, and the like. Statistical tools such as box plot statistics tools can also be used to process the evaluation results of individual patents in the group in order to obtain more accurate and reasonable evaluation results for the patent group.

The patent grading can be performed before or after the mathematical operation (ie, the calculation result is normalized into the score within the preset range as described above), and if it is the former, the value of each patent is first graded, and then Mathematical operations are performed on the value scores of each patent to obtain the evaluation results for the patent group, and then the evaluation results for the patent group are rounded; if the latter, the mathematical results of each patent evaluation result are first obtained. The evaluation result of the patent group is then normalized into the scores within the preset range, that is, the value of the patent group is graded.

FIG. 6 shows a schematic diagram of a comprehensive evaluation result of patent value according to an embodiment of the present disclosure. As shown in Fig. 6, in order to enable the user to obtain the evaluation result of the value more intuitively and quickly, and to realize the screening and statistics of the patent value, the above comprehensive evaluation scores are normalized. For example, according to the distribution of the comprehensive evaluation scores of patents in the whole library, the comprehensive evaluation scores of patents are segmented according to the 10-point scoring mechanism to ensure that the patent data of each score is in the whole library. The proportion of the distribution is distributed according to certain rules, and the probability distribution of the patent value decreases from 10 points to 1 minute. Similarly, for each category of evaluation scores, the same method can be used for 10-point processing.

In addition, the model in this embodiment is an evaluation model for the comprehensive value design of Chinese invention and utility model patents. Although the obtained patent value reflects the probability distribution of patent value as a whole, if the core influence factor of patent value is to be quickly judged The comprehensive evaluation scores of the three evaluation dimensions of the patent value can also be hierarchically designed in a manner similar to the above. On the basis of obtaining the comprehensive value of the patent, the user can comprehensively understand the specific value of the patent based on the comprehensive score of the value of each dimension displayed in the list of relevant evaluation basis.

Compared with inventions and utility models, Chinese design patents have different objects of protection. There are differences in parameter selection and value judgment. These differences can be considered in model construction and evaluation, but the basic design ideas and methods are the same. , no longer repeat them here.

Although some example implementations for optimizing the patent value evaluation model are given above, the embodiments of the present disclosure are not limited thereto, and the optimization of the model will be further described in detail below. The above embodiment introduces the establishment process of the patent value evaluation model for the actual model of the comprehensive value evaluation of Chinese high-value patents. In the above embodiments, the user needs to quickly identify the high patent value in the retrieval process, and fully utilize the patent document information and relevant patent laws and operational information that have been published since the implementation of the patent system in China in 1985, in a large number of experiments and analysis. On the basis of this, based on the invention patents granted by China, the basic model of China's authorized invention patents was established. On this basis, the design adjustments were made for Chinese patent applications, Chinese utility models and Chinese design patents. Objective evaluation and grading of the comprehensive value of Chinese patents. From the actual verification effect of the model, this model can achieve the desired recognition and resolution effect on Chinese high-value patents.

In addition, in the search process, in addition to hoping to quickly identify high-value patents in China, users often hope to simultaneously classify the value of patents in other countries and regions and display them simultaneously in the search results. To this end, the parameters of the model and the weighting of the models can be adjusted according to the characteristics of patents in different countries and regions, so that they can be integrated with the Chinese patent value evaluation model to form a complete system for the evaluation of patent comprehensive value.

On this basis, more detailed and personalized model design can be carried out. For example, for the special database in the field of chemical materials in China, after the completion of the basic database, high-value patent samples can be selected and analyzed to form patents in professional fields. Value assessment model.

Specifically, patent information can be applied to technical intelligence analysis, novelty analysis of technology or patents, free implementation (FTO) analysis, competitive intelligence analysis, patent operation analysis, anti-infringement analysis, technical cooperation, talent introduction and other scenarios. The purpose of different application scenarios is different, and the understanding or demand for “high-value patents” is different. Therefore, on the basis of the above comprehensive evaluation model, we need to adjust the corresponding parameters and weights in the model for different application scenarios to form a refined value evaluation model.

When focusing on the contribution of a parameter to the patent value, the weight of the parameter can be increased in the evaluation formula, and vice versa. The following is a brief description of the different application scenarios and the focus of the patent value evaluation model.

(1) Patented technology

The patent technology value analysis is mostly used for research activities aimed at acquiring technical solutions and their technological evolution processes and development trends, such as basic research and analysis of special technologies, technical research before project establishment, and technical solutions during project establishment. Scenes related to technical solution query in project development. In such application scenarios, the focus of users' attention is on the innovation of the technology solution itself and its origin, evolution, and application, especially those technologies that have not been of long origin, high novelty, rapid development, and high influence. patent. Therefore, in the design of the patent value evaluation model for the purpose of designing technical information, the direct parameters should focus on the patent application date (patent age), novelty, type and authorization of patents, average age of cited documents, and number of IPC classifications. Number of citations of patents or patent families (especially the number of patents cited by him) and the number of IPC classifications involved, the number of IPC classifications involved in citation patents, the number of subordinate claims corresponding to independent claims, the number of specifications and the number of inventors, etc. Indirect parameters should focus on the applicant's comprehensive technology advancement index, the technical relevance of the industry, the technology emerging degree, the technical heat and the technology life cycle.

(2) Analysis of infringement risk

The infringement risk analysis includes patent infringement retrieval analysis for specific technical solutions and early warning analysis of patent infringement risks for specific technical fields.

Patent infringement search is also called patent infringement investigation, that is, the process of identifying a patent that may have a risk of infringement by searching for a specific technical solution, and analyzing whether the existing technical solution falls within its protection scope. The patent infringement search also belongs to the micro-search for specific technical content. The main content of the patent infringement is whether there is a patent similar to the technical content of the technical solution being analyzed. Whether the current legal status of the patent is valid, if effective, relevant technology Whether the program falls within its scope of protection.

Although the infringement search is also a non-directional micro-search for specific technical solutions, the content of the infringement analysis involves the protection of patent rights, the quality of patent writing, the scope of protection of rights, the nature of rights subjects, and the protection of patents and rights protection strategies. Etc., it will affect the probability of patent infringement risk. Therefore, when conducting infringement search, it is necessary to distinguish the risk probability of patent infringement risk, that is, it is necessary to design a set for the purpose of infringement search. Patent risk value assessment model. In the design of direct parameters, the model pays more attention to the validity of current patents related to legal factors, the type of patents, the number of independent claims, the number of dependent claims, the number of independent claims or technical features, the number of manual pages, patent litigation And the case of the same family patent litigation, the licensing and transfer of patents, the invalidation of patents and family patents, the authorization of the same family patents, etc. Indirect parameters should pay more attention to the applicant's technical relevance, comprehensive patent application quality index, and applicant's comprehensive attack. Sex index and so on.

Patent risk warning refers to the process of patent search and analysis for the technical field of interest in order to detect or circumvent possible infringement risks in advance, identify patents that may be at risk, and classify, monitor and take countermeasures. Unlike patent infringement retrieval, patent risk warnings are not specific technical solutions, but all technologies related to the technology, product or business of interest. However, as with infringement retrieval, the focus of attention is on the protection of rights of related patents and the probability of occurrence of risks. The patent value analysis model of patent risk warning can refer to the model of patent infringement retrieval.

(3) Analysis of patent operations

The purpose of patent operation analysis is to quickly find high-value patents with operational value. Therefore, we should first pay attention to the operation of patents, namely, licensing, transfer, pledge and other activities. Among them, patent licensing and transfer are divided into internal transfer. And the licensing and licensing and transfer outside the organization, the former is more focused on the internal strategic management of the organization, while the latter is more objective to reflect the market operating value of the patent. Because patent litigation, invalidation and other activities can also reflect the impact of patents on the market from another aspect, and thus reflect its operational value, it should also be our focus. In addition, the citation of patents reflects the technical influence of patents, while patents that have significant technical impact are often patents with relatively large operational value; in addition, the validity of patents, the scope of protection, the operation of the same family of patents, and the invalidity of litigation And the applicant's comprehensive operational capability index, etc., also has an important reference role for the operational value of the patent, and can be used as an important index of the patent operation value analysis model.

(4) Quality control of patent applications

Patent value is a comprehensive manifestation of technical value, legal value and market value. The technical value is the basis for the formation of high-value patents, while the patent application is the key to transforming the technical value into legal value. The combination of technical value and legal value becomes the support of market value. Therefore, the management and monitoring of the quality of patent applications is also one of the focuses of the public. The patent application quality evaluation model can be used to examine the quality of the applicant's own or competitors, partner patent applications, and in particular to examine the quality of service of patent agencies.

When constructing the patent application quality value evaluation model, the number of independent claims, the number of dependent claims corresponding to independent claims, the number of independent claims (or technical features), the number of manual pages, the number of drawings, etc. For direct parameters related to patent writing, if possible, refer to the average indicators of the above indicators in the corresponding technical fields. Applicant's comprehensive patent application quality index, patent agency comprehensive patent application quality index, and indirect parameters are all related to the quality of patent applications, but they are comprehensive evaluation indicators of patent application quality. These indicators are used for comprehensive value analysis or competitor analysis. In the case of operational analysis, it can be used as a reference for the patent quality of related parties, but it is not recommended to give excessive weight when evaluating the quality of specific patent applications.

(5) invalid patent

A patent is a legally exclusive right granted to a patentee by law, but such a right is conditional. First of all, patent protection is regional. After a patent is filed abroad, if it does not apply for protection in China within the time prescribed by law, it will lose its priority in China. If the application has been disclosed before, the patent corresponds to The technical solution will lose the opportunity to be authorized in China; secondly, after the patent application, it needs to be reviewed before authorization. If the patent is rejected or withdrawn after the disclosure, the applicant will also lose the monopoly right; again, after the patent is granted, the application The person must pay the annual fee on time to maintain normal rights, and the patent protection period is limited. Once the protection period expires, the applicant will lose the corresponding rights; in addition, the authorized patent may be challenged by others, if it is all Or partially invalid, the patentee will also lose all or part of the rights. These patents that lose their rights will enter the public domain in the country and be freely available to the public. We refer to the above several situations as invalid patents. (Note: It should be noted that the loss of the rights of an applicant or a right holder does not mean that the patent must become a public technology. The public should also pay attention to whether there is any The corresponding right of the applicant or other person to be legally and effectively owned; in view of the fact that the patent protection is territorial, the invalidation of a patent in one country or region does not mean that its protection in other countries or regions also loses its effectiveness).

In some application scenarios (such as patent infringement risk analysis and patent operation analysis), invalid patents have little value. On the other hand, the purpose of the patent system is to enable the public to obtain the fruits of technological innovation and to promote social progress through application and improvement. A large number of invalid patents are a free treasure trove of knowledge that the public can easily access. Invalid patents can be directly used without risk when it is determined that there are no other legal rights, which greatly saves research and development time and R&D investment, and its value is also understandable. Therefore, it is of great significance to specifically classify and evaluate the value of the use of invalid patents.

The value evaluation of invalid patents and patent operation analysis are quite different in terms of application purposes and scenarios. However, except for the fact that the patents currently have the same judgment conditions as to whether the target country is invalid, the other principles for high-value patent judgment are basically the same. The value evaluation model can be obtained by slightly modifying the patent operation value evaluation model.

It is particularly emphasized that the optimization of the above-mentioned value evaluation model is only to increase the weight of one or some parameters in the patent value contribution for the user's actual application scenario, and the patent priority levels are separately set to facilitate Quick search browsing should not be understood as a model for patent screening or exclusion. In view of the complexity and uncertainty of the patent document information and the patent value itself, any patent value evaluation system formed by the mathematical model established by the patent's own parameters is to evaluate the probability of occurrence of the relevant event, rather than the absolute judgment of the value of the patent itself. in conclusion. In addition, the above models can be used alone or in combination with each other. For example, the patented value analysis of the search results can be carried out first, and then the operational value analysis can be carried out. The results of the two can be combined to quickly find out the patents with higher operational value and patented technology value.

In addition, the adaptive total and refined design of the above patent value evaluation model are designed according to the patent type, application area and application scenario of different objects, which makes the evaluation result more targeted and therefore more scientific and reasonable.

In the patent value evaluation model mentioned above, the selection of each parameter and the setting of its weight are based on the analysis of the sample data, and the corresponding analysis assignment method is used. On the whole, the results of evaluation and grading are consistent with the law of probability distribution of high-value patents, and thus can satisfy most commonly used retrieval and analysis scenarios. Usually, once a value evaluation model is determined, except for occasional local optimization adjustments, these parameters and their weights are generally fixed, and the user has only the right to use, and has no right to choose adjustments freely. Although such a value evaluation model can meet the requirements of general search and analysis, for higher-end users, it is necessary to personalize the model for different application scenarios and different needs, and realize flexible design of the model.

(1) Flexible design of patent categories for different characteristics

The development of technology is unbalanced. For example, in different periods of the same technology field, technical heat, technology life cycle, market competition, etc. may vary greatly. If the same model is used to analyze the value of patents, it will appear. Biased. At this time, the user can be allowed to input the search conditions, specify the basic data and the high-value patent samples therein, respectively calculate the corresponding patent parameters, compare these parameters with the original model, and then re-assign new weights to form a targeted A more scientific and accurate value evaluation model in the technical field.

Similarly, it is also possible to analyze the characteristics of different applicant types (such as enterprises, colleges, individuals, etc.), patents applied in different periods, and even patents applied by different applicants, so as to form a corresponding personalized value evaluation model. .

(2) Flexible design for individualized needs of different users

The needs of users are varied, no matter how perfect the curing model can not adapt to different individual needs of users. For example, if users are more concerned about emerging patent technologies when analyzing the value of patented technology, they may wish to increase the weight of the patent application date and novelty; if they are more concerned with the influence of technology, they hope to increase The weight of the cited case of a patent or patent family. Therefore, on the basis of the existing value evaluation model, allowing users to adjust the corresponding indicators and realize the flexible design of the patent value evaluation model will make the model more humanized and flexibly adapt to different kinds of Application scenario.

The characteristics, limitations and applicable application scenarios of the patent value evaluation model

(1) Objectivity

The indicators used in the evaluation model are all objective parameters extracted from the patent information. The weight of each parameter's influence on the value is calculated based on the results of the big data analysis of the current Chinese patent database. The patent value reflected by the patent value is the patent. The actual expression of each objective parameter is basically not affected by human factors, except for minor adjustments based on actual conditions in the later iteration stage.

(2) Comprehensiveness

In the process of constructing the evaluation model, the various parameters that may affect the patent value are tested and verified comprehensively. The various dimensions of patent value are comprehensively analyzed. The various influences of patent parameters are fully considered in the process of model construction. Factors, the final evaluation results are the comprehensive application of multiple parameters, the effect is significantly better than the single parameter evaluation.

(3) Real-time

The evaluation indicators in the evaluation model are dynamic indicators collected in real time in the database, and are updated in real time with the announcement of patents, changes in legal status, and the appearance of new citation information. The patent value model itself will be adaptively adjusted according to the changes of the database. To ensure that the current value of the patent is truly reflected.

(4) Accuracy

Based on the world's advanced credit card value evaluation model, the evaluation model is designed according to the characteristics and laws of Chinese patent information, which can truly reflect the value distribution of Chinese patents. Its accuracy has been verified in long-term practical applications.

(5) Convenience

The weights of the various parameters in the evaluation model are automatically calculated by the computer. In the case of selecting high-value patent samples, the participation of experts is not required (especially without the need for experts to separately evaluate the indicators for each parameter). The construction of the model is convenient and fast.

(6) Flexibility

The evaluation model is based on the comparison between the selected high-value patent sample parameters and the average parameters of the database patent. Therefore, the user can select different analysis targets and high-value patent samples according to actual needs, and quickly construct the corresponding model. For example, users can use the Chinese invention and utility model patent applications of the applicant type after the year 2000 as the basic data, and select some high-value patent samples from them to construct the value evaluation of the patents of Chinese universities after 2000. Models, and objectively compare and analyze the value of these patents.

The weight of each parameter's influence on the value of the model depends on the distribution of the relevant parameters of the high-value patent sample. The selected high-value patent sample is different, and the weight of the parameter in the model will also be different, which will lead to the patent value. The results of the evaluation vary. Therefore, the strategy of selecting high-value patent samples is a key factor in the construction of the value model. For the weight adjustment mentioned in the above embodiments, an appropriate implementation manner may be selected according to requirements, for example, it may be performed manually or automatically.

The patent value evaluation model of the embodiment of the present disclosure is based on the industry's accepted standard for evaluating the patent value, using the statistical analysis results of the big data of the Chinese patent database, and the value evaluation model for the tailor-made design of the characteristics of the Chinese patent information. The patent value can be screened and sorted according to the distribution probability of high-value patents, which enables users to quickly focus on high-value patents in a large number of patents, and significantly improve the efficiency and effectiveness of patent information utilization.

The embodiments of the present disclosure quickly understand the important information hidden in the patent by analyzing the quantity and distribution law of the high-value patents in the patent search results, thereby more accurately and comprehensively grasping the overall situation of the analysis object, and further formulating relevant decisions or Provide a strategic reference for response measures. For example, it is possible to screen and analyze a company's high-value patents, obtain the number and trend of their applications, the distribution of technology fields, technical impacts, application countries, inventor inputs, litigation, invalidation, licensing and transfer characteristics. Its intelligence in terms of technology and market competitiveness, development strategy, core talents and offensive and defensive characteristics, and then objectively assess its competitive threats to its own market, or find opportunities for cooperation in a timely manner, and develop corresponding Respond or take appropriate action.

In the aspect of value analysis for a specific patent, the embodiment of the present disclosure enables the user to quickly understand the patent value level of each relevant dimension of the patent by viewing the detailed information of the patent value on the basis of generally understanding the comprehensive value of the patent. The value, as well as the details of each important parameter affecting it, quickly determine the actual value of the patent.

The verification results of the embodiments of the present disclosure are described below with reference to FIGS. 7A through 7C.

In the construction of the patent value evaluation model, the evaluation effect can be verified through the actual application scenario. In this embodiment, a high-value patent is selected to verify the patent value evaluation model. In order to ensure the objectivity of the verification result, when selecting the data sample of the effect verification, avoid coupling with the establishment rule of the model, that is, not using the high-value patent sample selected when the model is established as the verification sample, nor the relevant parameters in the model. A search is performed to obtain a verification sample. In addition, to ensure that the selected high patent value should have a certain degree of credibility, for example, the patent value has been verified by the market or widely recognized by the industry. In the present embodiment, it is considered to select a verification sample from a sample selection criterion different from that at the time of model establishment.

FIG. 7A is a graph showing the value analysis of the patents of the first to the eighteenth China Patent Awards by using the patent value evaluation method of the embodiment of the present disclosure. The China Patent Award is a joint evaluation activity between the State Intellectual Property Office of China (SIPO) and the United Nations World Intellectual Property Organization (WIPO). It is also the only government department award in China that rewards inventions and inventions for granting patents. It has been successfully held so far. Eighteenth. The award criteria not only emphasize the patent technology level and innovation level of the project, but also pay attention to its application in the market transformation process, and also put forward requirements for its protection status and management. The awards are recommended by local IP Offices, State Council units, state-level industry associations, or academicians of the Chinese Academy of Sciences/Chinese Academy of Engineering. They are selected by the IP Office and senior experts from various industries, and have high credibility. Can be used as a verification sample for the value model.

A total of 3,175 invention and utility model patents were obtained from the first to the 18th China Patent Awards (including the China Patent Award and the China Patent Gold Award). After grading the patent value, the derived analysis chart is shown in Fig. 7A. As can be seen from Figure 7A, there are 871 patents with a full value of value, accounting for 27.4% of the total sample (only 4.5% of patents with a score of 10 in ordinary patents), and 2664 patents with a value greater than or equal to 8. Pieces, accounting for 83.9% of the total number of samples, and the total value of patents less than or equal to 5 is only 127, accounting for only 4% of the total number of samples (the percentage of patents with a patent value of 1 to 5 in ordinary patents is 55.5%). By looking at the current legal status of patents, these 127 patents have all expired at present (according to the selection criteria of the Chinese Patent Award, these patents should all be authorized and valid at the time of selection). The above cases fully demonstrate that the patent value model we are currently building has a very good match with the results of the China Patent Awards and can be used to screen high-value patents.

Of course, there are also some so-called medium-value patents with patent value of 6, 7, and 8. Some of the reasons are due to the partial patent invalidation, which leads to the reduction of patent value. At the same time, it should also consider the selection criteria of the Chinese Patent Award. It includes economic benefits, market transformation, and other factors that cannot be predicted by relying solely on the patent value model. On the other hand, the patent value evaluation model is not an “artifact” to identify whether a specific patent is absolutely a high-value patent. Although it cannot guarantee a hundred patents, its probability of judging high-value patents is quite reliable. We are a competent assistant in identifying high-value patents at work.

FIG. 7B is a graph showing the value analysis of the ETSI Chinese standard essential patent by the patent value evaluation method of the embodiment of the present disclosure.

Regarding Standards-Essential Patents (SEP), the International Telecommunication Union (ITU) defines it as "any patent or patent application that may cover, in whole or in part, the draft standard." The standard essential patents have significant application value because they may cover the draft standard in whole or in part. We selected the standard essential patents (6851 pieces) related to the ETSI (European Telecommunications Standards Institute) standard for patents for inventions and utility models in China, and analyzed the current value distribution as shown in Figure 7B. Show. As can be seen from Figure 7B, there are 6,129 patents with a patent value greater than or equal to 9, accounting for 89.5% of the total number of samples, and only 151 patents with a value of less than or equal to 7, accounting for 2.2% of the total number of samples, and these patents have 64.9% is a patent that has expired.

FIG. 7C illustrates the evaluation of the comprehensive value of a single patent and the evaluation of the patent price of each dimension by the patent value evaluation method according to an embodiment of the present disclosure. In the example of FIG. 7, the patent publication number is CN102868498B, and the patent name is "codebook generation method, data transmission method and device". The technical stability and technological advancement are extracted from the patent with the patent value equal to 10. The key parameters and scoring conditions of sex and protection range are shown in Figure 7C. As can be seen from Figure 7C, the indicators displayed in the list are only some of the common parameter indicators used in the patent value evaluation model, but the “standard essential patent” factor is not taken into account under the premise that the patent value evaluation model affects the parameters. Can be automatically identified as a high-value patent, indicating that the model can reasonably use the big data statistical information of each patent parameter in the construction process, and scientifically and rationally combine it to form a credible patent value evaluation system. .

From the previous high-value patent grading effect verification experiment, it can be seen that the patent value model of the embodiment of the present disclosure objectively has a very significant effect on quickly identifying high-value patents. Below we use this model to calculate the statistics of Chinese patent operations, litigation and other data, and learn some of the distribution rules.

The number of patent applications (including invention applications, utility models and designs, excluding authorized invention texts) in the Chinese patent database is counted, and the proportion of each legal event occurs, as shown in Table 2 below.

Table 2

As can be seen from Table 2, high-value patents with a patent value of 9 or 10 account for less than 10% of all patents, but contain more than half of the patents that have been litigated, licensed or pledged, and 40%. The above-mentioned patents that have undergone review or customs filing, and more than 30% of patents that have been transferred or invalidated. Further analysis of the above-mentioned patents with low patent value (patent value less than or equal to 4) found that most of them were rejected or totally invalid, patent stability was poor, and protection scope was narrow. Therefore, the patent value model can help users effectively screen high-value patents.

The probability of occurrence of various legal events for patents of various values is further analyzed below with reference to Table 3:

table 3

It can be clearly seen from Table 3 that the probability of legal events such as litigation, pledge, transfer, license, customs filing, review, invalidation, etc. of patents with high patent value is far greater than that of patents with lower value.

The application of patent value in patent retrieval and analysis practices is briefly illustrated below with reference to specific examples with reference to Figures 8A and 8B.

FIG. 8A shows a schematic diagram of a user interface for patent value evaluation using a patent value evaluation method according to an embodiment of the present disclosure. Assume that the enterprise that produces soybean milk machine intends to understand the important patents of soymilk in this industry. You can search for the patent containing the keyword “Soya-bean milk machine” in the title or abstract, and choose the sorting according to the patent value in the search results, as shown in Figure 8A. Shown. In the first place in Figure 8A is a utility model patent "a double-layer under-cover soymilk machine" (application number: CN201414387Y). As can be seen in the label next to the patent, this patent has experienced various legal events such as litigation, transfer, licensing, customs filing and invalidation review. Further investigation of this patent found that the patent was a BBK patent purchased by Jiuyang Co., Ltd. In the subsequent patent infringement lawsuit, Jiuyang Company won the case and compensated 5.4 million yuan. Small utility models do not allow inventions and play a major role in patent rights protection.

Next, we analyze the distribution of the value of the applicant's patent.

FIG. 8B shows a schematic diagram of the applicant's patent layout using the patented value evaluation method according to an embodiment of the present disclosure. As can be seen from Figure 8B, among the Chinese patents applied by Jiuyang Co., Ltd., there were 4, 19, 21, 6, 6 and 8 cases involving litigation, transfer, licensing, customs filing, review and invalidation; The patents of Chinese patents applied by Jiuyang Company are equal to 9 or 10 patents, and 4, 8, 20, 3, 4 and 8 cases have been found in the above legal incidents, including most of the above-mentioned laws. An important case of the incident. It can be seen that the applicant has more high-value patents.

The above is only a simple indication of the application of the patent value model. In practice, the patent should be carefully searched according to the needs, and the patent value model should be used to screen the search results, combined with the actual content and technology of the patent and the application of the market. Conduct in-depth, specific analysis.

Embodiments of the present disclosure also provide a computer readable storage medium for storing instructions that, when executed by a processor, cause a processor to perform the above method.

By way of example, embodiments of the present disclosure may also be described in the context of machine-executable instructions, such as in a program module that is executed in a device on a real or virtual processor of a target. Generally, program modules include routines, programs, libraries, objects, classes, components, data structures, and the like that perform particular tasks or implement particular abstract data structures. In various embodiments, the functionality of the program modules may be combined or divided between the described program modules. Machine-executable instructions for program modules can be executed within a local or distributed device. In a distributed device, program modules can be located in both local and remote storage media.

Computer program code for implementing the methods of the present disclosure can be written in one or more programming languages. The computer program code can be provided to a general purpose computer, a special purpose computer or a processor of other programmable data processing apparatus such that the program code, when executed by a computer or other programmable data processing apparatus, causes a flowchart and/or block diagram. The functions/operations specified in are implemented. The program code can execute entirely on the computer, partly on the computer, as a stand-alone software package, partly on the computer and partly on the remote computer or entirely on the remote computer or server.

In the context of the present disclosure, a machine-readable medium can be any tangible medium that contains or stores a program for or relating to an instruction execution system, apparatus, or device. The machine readable medium can be a machine readable signal medium or a machine readable storage medium. A machine-readable medium can include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination thereof. More detailed examples of machine readable storage media include electrical connections with one or more wires, portable computer disks, hard disks, random access memory (RAM), read only memory (ROM), erasable programmable read only Memory (EPROM or flash memory), optical storage device, magnetic storage device, or any suitable combination thereof.

The above description is only a preferred embodiment of the present disclosure, and is not intended to limit the disclosure, and various changes and modifications may be made to the present disclosure. Any modifications, equivalents, improvements, etc. made within the spirit and scope of the present disclosure are intended to be included within the scope of the present disclosure.

Claims (21)

1. A method for constructing a patent value evaluation model, which comprises:

   Select parameters that affect the value of the patent;

   Selecting patent data samples according to the preset patent value evaluation criteria;

   Based on the selected patent data samples, an evaluation formula reflecting the relationship between the selected parameters and the patent value is established.
2. The method of constructing a patent value evaluation model according to claim 1, wherein the parameter comprises at least one of the following:

   Parameters related to the type, status, content, and review process of the patent;

   Parameters related to the citation and citation of the patent;

   Parameters related to the layout of the patent;

   Parameters related to the maintenance of patents;

   Parameters related to the inventor and applicant of the patent;

   Parameters related to the operation of the patent;

   Parameters related to patent disputes and disputes.
3. The method for constructing a patent value evaluation model according to claim 1, wherein the selecting parameters affecting the patent value includes:

   Selecting factors that affect the value of the patent and sub-parameters used to evaluate the factors;

   The factors are evaluated based on the sub-parameters, and the evaluation results are taken as the selected parameters.
4. The method for constructing a patent value evaluation model according to claim 1, wherein the method further comprises: after selecting the patent sample, establishing a selected parameter and a patent value based on the selected patent data sample. Relationship map; and

   The evaluation formula is established based on the relationship map.
5. The method for constructing a patent value evaluation model according to claim 1, wherein the selecting a patent data sample according to a preset patent value evaluation standard comprises:

   Selecting a patent that meets the preset first condition as the first set of patent data samples;

   A patent that meets the preset second preset condition is selected as the second set of patent data samples.
6. The method for constructing a patent value evaluation model according to claim 5, wherein the second group of patent data samples are selected from the first group of patent data samples, and the first condition includes the patents recorded in the country. All inventions and utility models in the database, the second condition comprising at least one of the following:

   a patent infringement lawsuit and a successful patent;

   Maintain the payment of the patent until the expiration of the patent period;

   a patent that is put into operation and receives returns;

   a Chinese patent that has been requested to be invalid and has been sentenced to remain valid;

   Forming industry-standard patents;

   a patent that has been used more than a preset number of times;

   Patent applied for in multiple countries.
7. The method for constructing a patent value evaluation model according to claim 5, wherein the evaluation formula for establishing a relationship between the selected parameter and the patent value based on the selected patent data sample comprises:

   Calculating the distribution of the selected parameters in the first set of patent data samples;

   Calculating the distribution of the selected parameters in the second set of patent data samples;

   Calculating a comparison between the distribution of the selected parameters in the first set of patent data samples and the distribution of the selected parameters in the second set of patent data samples;

   The relationship between the selected parameter and the comparison result is fitted based on the comparison result as an evaluation formula reflecting the relationship between the selected parameter and the patent value.
8. The method for constructing a patent value evaluation model according to claim 1, wherein the evaluation formula for establishing a relationship between the selected parameter and the patent value based on the selected patent data sample comprises:

   Extracting the patent corresponding to each value of the parameter;

   Patent value evaluation for each patent corresponding to each value, and based on the evaluation result, the influence of the value of the parameter on the patent value is set;

   Based on the influence of each value of the parameter on the patent value, the evaluation formula that reflects the relationship between the parameter and the patent value is obtained.
9. The method for constructing a patent value evaluation model according to claim 8, wherein the patent data sample comprises a patent data sample of a single patent or a patent data sample of a composite patent converted from a plurality of patents.
10. The method for constructing a patent value evaluation model according to claim 1, wherein there are a plurality of selected parameters, and the relationship between the selected parameters and the patent value is established based on the selected patent data samples. The evaluation formula includes:

    For each selected parameter, a sub-evaluation formula embodying the relationship between the parameter and the patent value is established based on the selected patent data samples;

    The weighted sum of the sub-evaluation formulas is taken as an evaluation formula for the selected overall parameters.
11. The method for constructing a patent value evaluation model according to claim 1, wherein the selected parameters are divided into a plurality of groups, and the selected patent data sample is established based on the selected parameter and the patent value. The evaluation formula for the relationship includes:

    For each selected group of parameters, a sub-evaluation formula reflecting the relationship between the set of parameters and the patent value is established based on the selected patent data samples;

    The weighted sum of the sub-evaluation formulas is taken as an evaluation formula for the selected overall parameters.
12. The method for constructing a patent value evaluation model according to claim 1, further comprising:

    Receive other parameters;

    Based on the other parameters, at least one other model for evaluating the patent value is established, and the other models are combined with the established evaluation formula to obtain a finally constructed patent value evaluation model.
13. A computer readable storage medium for storing instructions that, when executed by a processor, cause a processor to perform the method of constructing a patent value evaluation model according to any one of claims 1 to 12.
14. A patent value evaluation method, characterized in that it comprises:

    Construct a patent value evaluation model;

    Evaluation of the target patent based on the constructed patent value evaluation model;

    Optimize the patent value evaluation model based on the evaluation results of the target patent.

    Wherein, the construction patent value evaluation model includes:

    Select parameters that affect the value of the patent;

    Selecting patent data samples according to the preset patent value evaluation criteria;

    Based on the selected patent data samples, an evaluation formula reflecting the relationship between the selected parameters and the patent value is established.
15. The patent value evaluation method according to claim 14, wherein the constructed patent value evaluation model comprises a first model and a second model different from the first model, the evaluation result based on the target patent The evaluation model for optimizing patent value includes:

    a target patent to be evaluated by the first model and/or the second model as having a value in a desired range as a reference patent;

    The patent data sample used to construct the first model and/or the second model is updated using the reference patent.
16. The patent value evaluation method according to claim 14, wherein the constructed patent value evaluation model comprises a first model, and the evaluation model for optimizing the patent value according to the evaluation result of the target patent comprises:

    The target patent to be evaluated by the first model as having a value in the desired range as a reference patent;

    The patent data sample used to construct the first model is updated using the reference patent.
17. The patent value evaluation method according to claim 14, wherein the constructed patent value evaluation model comprises a first model and a second model different from the first model, and the patent value evaluation method further comprises:

    The pre-selected verification data is evaluated using the first model and the second model, respectively, to obtain a first evaluation result and a second evaluation result;

    Comparing the first evaluation result with the second evaluation result;

    The first model and/or the second model are adjusted based on the comparison results.
18. The patent value evaluation method according to claim 14, wherein the constructed patent value evaluation model comprises a first model, and the patent value evaluation method further comprises:

    The pre-selected verification data is evaluated using the first model to obtain a first evaluation result;

    Evaluating a patent data sample used in constructing the first model to obtain a second evaluation result;

    Comparing the first evaluation result with the second evaluation result;

    The first model is adjusted based on the comparison result.
19. The patent value evaluation method according to claim 14, wherein the patent value evaluation method further comprises:

    After evaluating the target patent according to the constructed patent value evaluation model, the evaluation result is normalized to be converted into an evaluation score value within a preset range.
20. The patent value evaluation method according to claim 14, wherein the target patent includes a patent group, and the evaluation of the target patent according to the constructed patent value evaluation model includes:

    Evaluate each patent in the patent group based on the constructed patent value evaluation model;

    Mathematical calculations were performed on the evaluation results of each patent in the patent group to obtain evaluation results for the patent group.
21. A computer readable storage medium for storing instructions that, when executed by a processor, cause a processor to perform the patent value evaluation method according to any one of claims 14-20.

Images