CN109002983B - Method and computer system for evaluating enterprise investment potential - Google Patents

Abstract

The invention is a method and computer system for evaluating enterprise investment potential, the method includes (1) collecting patent information of multiple enterprises; (2) providing a patent lead equation, which consists of a plurality of patent core indexes and corresponding weight coefficients thereof and is used for predicting financial information of a patent entity, wherein the patent lead equation generates a patent lead score which leads the financial information of the patent entity by a preset time lead period, and the patent lead equation is obtained by a construction method of the patent lead equation; (3) calculating data of corresponding patent core indexes of each enterprise based on the patent information of each enterprise; (4) generating a patent lead score of each enterprise through calculation of the patent lead equation based on the data of the patent core indexes of each enterprise; and (5) ranking the plurality of patent lead scores by a ranking procedure, the ranking result representing the investment potential of the plurality of enterprises.

Description

Method and computer system for evaluating enterprise investment potential

[ technical field ] A method for producing a semiconductor device

The invention is a divisional application with name of construction method and application of patent leading indicators in 201410283508.7 applied in 23/06/2014.

The invention relates to a digital content processing technology, in particular to a method for mining the relationship of patent information to enterprise financial information through the operation of the patent information so as to establish a prediction method and a related computer system of advanced financial information of the patent information.

[ background of the invention ]

With the rapid development of the scientific and technological industry and the emphasis on intellectual property rights, patents have been regarded as important indexes for industry or technology. For the right-minded people, the patent is not only a defense tool for protecting innovation and products, but also becomes the best attack weapon for competing with competitors on the industrial stage. Since companies have more patents, the more representative and influential they have in their competitive fields, the more important information they have become.

The World Intellectual Property Organization (WIPO) has reported that 80% of the technical content contained in the patent is not disclosed in other papers, journals and encyclopedias. The patent refers to the field of 'electric digital data processing'. The patent quantity and quality are superior units, and the energy and quality of the research and development are superior to those of other competitors. The patent has little effect on the market as it has legally exclusive rights. Therefore, for the scientific and technological enterprises based on the technical research and development, the patent quantity and quality are superior, and the product sales and performance are also superior to a certain extent.

In the prior art, many papers and researches indicate that the amount information of patents leads the sales amount information of products, and has a leading effect on explaining the development condition of the market, such as the following documents:

(1)Griliches,Z.(1990),Patent statistics as economic indicators：asurvey,Journal of Economic Literature,28(4),PP.1661-1707.

(2)Ernst,H.(1995),Patenting strategies in the German mechanicalengineering industry and their relationship to firm Performance,Technovation,5(4),PP.225-240.

(3)Adams,K.,D.Kim,F.L.JoLz,Trost,R.P.and Mastrogianis,G.(1997),Modeling and Forecasting U.S.Patent Application Filings,Journal of PolicyModeling,19(5),PP.491-535.

(4)Ernst H.(1997),The Use of Patent Data for TechnologicalForecasting：The Diffusion of CNC-Technology in Machine Tool Industry,SmallBusiness Economics,9(4),PP.361-381.

(5)Ernst H.(2001),Patent applications and subsequent changes ofPerformance：evidence from time-series cross-section analyses on the firmlevel,Research Policy,30(1),PP.143-157.

the above prior art indicates that patent information is more important and more interesting than other market information, so that a method for predicting market information by using patent information is an important research topic in investment and enterprise evaluation.

In the prior art, US patents US6175824 and US6832171 first proposed the assessment of marketing company financial performance with patent information. For the scientific and technological stock with more patents of listed companies, the united states patents US6175824 and US6832171 analyze the correlation between the Patent Index (PI) of the listed company in the past year and the Financial Index (FI) of the past year through a multiple regression analysis model, finally derive an evaluation equation based on the Patent Index (PI), calculate the scientific and technological value by the evaluation equation, compare the Market-to-Book Ratio with the innovation value, if the scientific and technological value is greater than the Market-to-Book Ratio, the investment potential is considered, if the scientific and technological value is less than the Market-to-Book Ratio, the investment potential is considered, and the investment potential is used as a tool for investment stock selection. The patentee CHIResearch further pushes the evaluation equation based on the patent information to the public, and accordingly establishes a special business profit model.

The prior art US6175824 and US6832171 are unique, but have the disadvantage that the equation for analyzing the Patent Index (PI) and the Financial Index (FI) is a typical multiple regression analysis model, the independent variable is the Patent Index (PI) in a time series, and the dependent variable is the Financial Index (FI) in the same time series. The multiple regression equation thus established, when the Patent Index (PI) at a certain time point is input, the output result is the corresponding Financial Index (FI) at the time point, not the Financial Index (FI) in the "future".

In the prior art, U.S. patents US6556992, US7657476 and US7716226 develop another set of evaluation method for innovation capability of listed companies by using a multiple regression analysis model based on Patent maintenance rate and other Patent Indexes (PI), and accordingly 300 listed companies are selected as sample stock and issued with global first Patent Index (OT300Patent Index) after weighted calculation. However, in the multivariate regression analysis models used in US6556992, US7657476 and US7716226, the independent variable Patent Index (PI) and the dependent variable Financial Index (FI) belong to the same time series, and although the independent variable is different from the prior art US6175824 and US6832171, the multivariate regression equation established by the multivariate regression models still obtains the corresponding Financial Index (FI) at a certain time point instead of the Financial Index (FI) in the future when the Patent Index (PI) at the time point is input.

It must be understood that the actual practice of investment equity selection, when invested by an investment institution, does not want to make a profit at the moment, but rather wants to make a profit when it is set at some point in time in the future. That is, when an investment institution invests, the current information that is desired to be grasped has a "prediction" effect on the profit of the "future", so that the investment risk can be reduced and the investment benefit can be ensured. The above prior art US6175824, US6832171, US6556992, US7657476 and US7716226 do not substantially have a "predictive" effect.

Therefore, for the financial data of the listed companies in China and the patent information thereof in China, if the relational model with the 'prediction' effect and the computer system for realizing the relational model are provided, the method not only contributes to the development of the technical strength of the analysis and utilization of the patent information, but also can promote the positive development of the investment method in the investment field, and has a positive supporting effect on the research, development and innovation of the industrial technology.

[ summary of the invention ]

Based on the foregoing drawbacks of the prior art, the primary objective of the present invention is to provide a method (100) for constructing a patent lead indicator, wherein the patent lead indicator (172) is used to predict financial information of a Patent Entity (PE), and the Financial Indicator (FI) of the patent lead indicator (172) leads the financial indicator (PE) of the Patent Entity (PE) by a predetermined time lead period (L), and the method (100) for constructing the patent lead indicator comprises:

(1) a plurality of Patent Entities (PE) and a plurality of Patent Indexes (PI) and Financial Indexes (FI) for describing each Patent Entity (PE) are set, and each Patent Index (PI) is calculated by patent information of each Patent Entity (PE).

(2) Setting a data collection period (121), wherein the data collection period (121) is composed of a time interval (T) and a time period number (N), and the time period number (N) is an integer not less than two;

(3) collecting patent index data (131) and financial index data (132) corresponding to each time interval (T) in a data collection period (121) of each Patent Entity (PE);

(4) combining a plurality of patent index data (131) and a plurality of financial index data (132) of a plurality of Patent Entities (PE) into first panel data (141);

(5) forming second panel data (151) which is normally distributed and standardized and divided by the first panel data (141) through a conversion operation program (152);

(6) setting a time lead period (L) and providing a first time series operation program (161) based on the time lead period (L), wherein the time lead period (L) comprises at least one time interval (T), the independent variable of the first time series operation program (161) is a patent index data (131) of the second panel data (151), and the dependent variable of the first time series operation program (161) is a financial index data (132) of the second panel data (151);

(7) setting a first threshold (171), calculating second panel data (151) through a first time series calculation program (161) and a time lead period (L), and screening at least one patent lead index (172) meeting the first threshold (171) from a plurality of Patent Indexes (PI).

The construction method (100) of the patent leading index provided by the invention is objective and rigorous, and is particularly suitable for Chinese patent information, including patent disclosure, patent invention authorization, utility model patent and appearance design patent; meanwhile, the method is suitable for mining the patent leading index (172) with leading enterprise financial information aiming at the patent information of other countries.

It is still another object of the present invention to provide a method (500) for constructing a patent lead equation (501) for predicting financial information of a Patent Entity (PE), the patent lead equation (501) generating a patent lead score (502), the patent lead score (502) leading the financial information of the Patent Entity (PE) by a predetermined time lead (L). The method (500) for constructing the patent lead equation (501) comprises:

(1) the second panel data (151) and the plurality of patent lead indicators (172) are obtained, and the second panel data (151) and the plurality of patent lead indicators (172) are obtained by the patent lead indicator construction method (100).

(2) Forming third panel data (521) by screening from the second panel data (151) based on the plurality of patent lead indicators (172);

(3) providing a second time series operation program (531) based on the time lead period (L), wherein the independent variable of the second time series operation program (531) is a plurality of patent lead indexes (172) of the third panel data (521), and the dependent variable of the second time series operation program (531) is a Financial Index (FI) of the second panel data (151);

(4) setting a second threshold (541), calculating third panel data (521) through a second time series operation program (531) and a time lead period (L), screening a plurality of patent core indexes (542) meeting the second threshold (541) from a plurality of patent lead indexes (172) and generating a patent lead equation (501), wherein the patent lead equation (501) is substantially composed of the plurality of patent core indexes (542) and corresponding weight coefficients (543).

It is another object of the present invention to provide a method (600) of assessing investment potential of an enterprise, comprising:

(1) collecting patent information (612) of a plurality of enterprises (611);

(2) providing a patent lead equation (501), wherein the patent lead equation (501) is obtained by the method (500) for constructing the patent lead equation, and the patent lead equation (501) is substantially composed of a plurality of patent core indexes (542) and corresponding weight coefficients (543);

(3) calculating data (631) of a patent core index (542) corresponding to each enterprise based on the patent information (612) of each enterprise (611);

(4) generating a patent lead score (502) for each enterprise (611) by calculation using a lead equation (501) based on data (631) of a patent core index (542) for each enterprise (611);

(5) the patent lead scores (502) of the enterprises (611) are ranked by a ranking program (651), and the ranking result (652) represents the investment potential of the enterprises (611).

The higher the lead score (502) of the above patent, the higher the data representing the corresponding Financial Index (FI) of the enterprise after the time lead period (L); the lower the patent lead score (502), the lower the value of the corresponding Financial Index (FI) of the enterprise after the time lead period (L). As the data of the Financial Index (FI) of the enterprise directly expresses the operation performance, the higher the value of the Financial Index (FI), the better the operation performance of the enterprise, and the more investment value. Since the patent lead score (502) represents the value of the corresponding Financial Index (FI) of the enterprise after the time lead period (L), the high and low ranks of the patent lead score (502) of the enterprise can select the objects with investment potential from the objects.

A further object of the present invention is to provide a computer system (700) for evaluating investment potential of an enterprise, comprising a patent information collecting device (710), an index calculating device (720), a patent lead score calculating device (730), and a score ranking device (740). The patent information acquisition device (710) collects patent information (612) of a plurality of enterprises; an index calculation means (720) calculates, based on patent information (612) of each enterprise, data (631) of a patent core index (542) corresponding to the enterprise; the patent lead score calculating device (730) generates the patent lead score (502) of each enterprise by calculating through a patent lead equation (501) according to the data (631) of the patent core indexes (542) of each enterprise, wherein the patent lead equation (501) is obtained by the construction method (500) of the patent lead equation, and the patent lead equation (501) is substantially composed of a plurality of patent core indexes (542) and corresponding weight coefficients (543); the score ranking device (740) ranks the plurality of patent lead scores (502), and the ranking result (652) represents the investment potential of a plurality of enterprises.

The method (600) and the computer system (700) for evaluating the investment potential of the enterprise are based on big data, objective operation and rigorous verification, are not only beneficial to the development of the technical strength of patent information analysis and utilization, but also can promote the positive development of the investment method in the investment field, and have positive supporting effect on the research, development and innovation of industrial technology.

Other specific features and non-obvious advantages of the present invention will be described in detail in the following sections.

[ description of the drawings ]

FIG. 1 is a flow chart of a method (100) for constructing a patent lead indicator according to a first preferred embodiment of the present invention;

FIG. 2 is a graph of the number and industry distribution of the analyzed parents of the companies listed on the Shanghai exchange;

FIG. 3 is an industry distribution of a sample of the analytics for a company listed on the Shanghai exchange;

fig. 4 is a schematic diagram of the first panel data (141);

FIG. 5 is a patent lead indicator (172) with a lead period of one year;

FIG. 6 is a patent lead indicator (172) with a lead period of two years;

FIG. 7 is a patent lead indicator (172) with a lead period of three years;

FIG. 8 is a patent lead indicator (172) with a lead period of four years;

FIG. 9 is a flowchart of a method 500 for constructing the lead equation according to a second preferred embodiment of the present invention.

Fig. 10A, 10B and 10C show a process of generating a patent core indicator (542) according to a second preferred embodiment of the present invention.

Fig. 11 is a flowchart of a method (600) for evaluating investment potential of an enterprise according to a third preferred embodiment of the present invention.

FIG. 12 is a diagram illustrating a computer system (700) for assessing investment potential of an enterprise according to a fourth preferred embodiment of the present invention.

[ detailed description ] embodiments

The present invention discloses a method (100) for constructing a patent lead index and applications thereof, wherein the basic knowledge of the patent information, the Patent Index (PI), the Financial Index (FI), etc. is known to those skilled in the relevant art, and therefore, the following description thereof will not be described in full. Meanwhile, the drawings referred to in the following description only express the schematic representation related to the features of the present invention, and are not necessarily drawn completely according to the actual size, which is described in the foregoing.

Referring to fig. 1, a first preferred embodiment of the present invention is a method (100) for constructing a patent lead indicator, in which the patent lead indicator (172) is used to predict financial information of a Patent Entity (PE), and the information of the patent lead indicator (172) leads the Financial Indicator (FI) of the Patent Entity (PE) by a predetermined time lead (L), the method (100) for constructing the patent lead indicator includes:

step 110: a plurality of Patent Entities (PE) and a plurality of Patent Indexes (PI) and Financial Indexes (FI) for describing each Patent Entity (PE) are set, and each Patent Index (PI) is calculated by patent information of each Patent Entity (PE).

Step 120: a data collection period (121) is set, the data collection period (121) is composed of a time interval (T) and a time period number (N), and the time period number (N) is an integer not less than two.

Step 130: and collecting the patent index data (131) and the financial index data (132) corresponding to each time interval (T) of each Patent Entity (PE) in the data collection period (121).

Step 140: a plurality of patent index data (131) and a plurality of financial index data (132) of a plurality of Patent Entities (PEs) are combined into first panel data (141).

Step 150: the first panel data (141) is converted into second panel data (151) which is normally distributed and normalized by a conversion operation program (152).

Step 160: setting a time-lead period (L) and providing a first time-series calculation program (161) based on the time-lead period (L), wherein the time-lead period (L) comprises at least one time interval (T), the independent variable of the first time-series calculation program (161) is the patent index data (131) of a Patent Index (PI) of the second panel data (151), and the dependent variable of the first time-series calculation program (161) is the financial index data (132) of a Financial Index (FI) of the second panel data (151).

Step 170: a first threshold value (171) is set, a first time series operation program (161) and a time lead period (L) are used successively, second panel data (151) are operated, and at least one patent lead index (172) meeting the first threshold value (171) is screened from a plurality of Patent Indexes (PI).

In the step 110, the Patent Entity (PE) is a rights body that has patent rights and can make a profit through patent rights operation, and is preferably a publicly issued company on the market, but not limited to the company on the market; the embodiment is also applicable to non-listed companies, and the rights main bodies sharing the equity and the equity income as long as the rights main bodies can accept external funds to enter the embodiments all belong to the application scope of the embodiment. The patents of the first preferred embodiment are not limited to issued patents, and may be any patents issued in the patent database, including patent publications, patent issued patents, utility model patents, design patents, and the like. Meanwhile, the method provided by the first preferred embodiment can effectively solve the problem that the information content of the Chinese patent is different from that of the United states patent, and is more suitable for all regional patents in the world.

In terms of the Patent Indices (PI), taking the chinese patent as an example, the Patent Indices (PI) include but are not limited to the following quantity indices that can be automatically computed by a computer, such as:

p1: total number of patents

P2: total number of patents issued

P3: utility model total number of patents

P4: total number of design patents

P5: total number of patents granted by invention

P6: mean patent life of invention disclosure

P7: mean patent life of utility model patent

P8: design patent mean patent life

P9: mean patent life of patent granted by invention

P10: mean period of examination of patent granted by invention

P11: number of patents on the current invention

P12: number of current utility model patents

P13: number of design patents in current date

P14: number of patents granted by current invention

P15: the average examination period of the patent granted by the invention at the current date is from the application date to the publication granting date

P16: total number of IPC classification numbers of current invention patent publication

P17: current utility model patent IPC classification number total

P18: total number of IPC classification numbers of current invention grant patent

P19: average number of IPC classification numbers of current invention patent publication

P20: mean number of the current utility model patent IPC classification number

P21: average number of IPC classification numbers of current invention grant patent

P22: total number of pages of current invention publication patent specification

P23: total number of pages of current utility model patent specification

P24: total number of pages of current invention granted patent specification

P25: average number of pages of current invention publication patent specification

P26: average number of pages in current utility model patent specification

P27: number of pages averaged over current patent specification

P28: total number of claims of current invention disclosure patent

P29: claim total number of current utility model patent

P30: total number of claims of the current invention granted patent

P31: average number of claims of current invention disclosure patent

P32: mean number of claims of current utility model patent

P33: average number of claims of current invention granted patent

P34: total number of independent rights of current invention disclosure patent

P35: the total number of the current utility model

P36: total number of independent rights of current invention grant patent

P37: number of independent rights of the current invention disclosure

P38: mean number of exclusive rights of current utility model patent

P39: independent average number of patents granted by current invention

P40: number of figures in the current patent publication

P41: number of figures of current utility model patent specification

P42: number of figures in patent specification granted to current invention

P43: average number of figures of current invention disclosure patent specification

P44: mean number of figures of current utility model patent specification

P45: average number of figures in patent specification granted to current invention

In the above Patent Index (PI), "current stage" means that the data operation of the Patent Index (PI) is limited to a specific time interval (T), for example, if the time interval (T) is year, P11 (number of currently-published patent patents) represents the number of all published patent patents within 1-12 months of a certain year; p15 (average examination period of patent granted during the current period, from application date to publication date) represents the average examination period of all patent granted during 1-12 months within a certain year. If the time interval (T) is season, P11 (number of published patents) represents the number of all published patents within 3 months of a certain season; p15 (average examination period of patent granted during the current period, from application date to publication date) represents the average examination period of all patent granted during 3 months in a quarter. And so on.

The 45 Patent Indexes (PI) can be further classified into patents for invention disclosure, patents for invention authorization, patents for utility model, and patents for appearance design, wherein the indexes for describing the patents for invention disclosure include:

p2: total number of patents issued

P6: mean patent life of invention disclosure

P11: number of patents on the current invention

P16: total number of IPC classification numbers of current invention patent publication

P19: average number of IPC classification numbers of current invention patent publication

P22: total number of pages of current invention publication patent specification

P25: average number of pages of current invention publication patent specification

P28: total number of claims of current invention disclosure patent

P31: average number of claims of current invention disclosure patent

P34: total number of independent rights of current invention disclosure patent

P37: number of independent rights of the current invention disclosure

P40: number of figures in the current patent publication

P43: average number of figures of current invention disclosure patent specification

The indexes used for describing the patent granted on the invention are as follows:

p5: total number of patents granted by invention

P9: mean patent life of patent granted by invention

P10: mean period of examination of patent granted by invention

P14: number of patents granted by current invention

P15: the invention of the current date authorizes the patent average examination period, from the application date to the issue date P18: total number of IPC classification numbers of current invention grant patent

P21: average number of IPC classification numbers of current invention grant patent

P24: total number of pages of current invention granted patent specification

P27: number of pages averaged over current patent specification

P30: total number of claims of the current invention granted patent

P33: average number of claims of current invention granted patent

P36: total number of independent rights of current invention grant patent

P39: independent average number of patents granted by current invention

P42: number of figures in patent specification granted to current invention

P45: average number of figures in patent specification granted to current invention

The indexes used for describing the utility model are

P3: utility model total number of patents

P7: mean patent life of utility model patent

P12: number of current utility model patents

P17: current utility model patent IPC classification number total

P20: mean number of the current utility model patent IPC classification number

P23: total number of pages of current utility model patent specification

P26: average number of pages in current utility model patent specification

P29: claim total number of current utility model patent

P32: mean number of claims of current utility model patent

P35: the total number of the current utility model

P38: mean number of exclusive rights of current utility model patent

P41: number of figures of current utility model patent specification

P44: mean number of figures of current utility model patent specification

The indexes for describing design patents are

P4: total number of design patents

P8: design patent mean patent life

P13: number of design patents in current date

The above-mentioned Patent Indicators (PI) from P1 to P45 are quantitative indicators describing the most sufficient characteristics of chinese patents, and as to which of the Patent Indicators (PI) have the effect of predicting financial information, it needs to be strictly analyzed and verified, which is the core of the present invention, and the following paragraphs will continue to describe the present invention.

In the aspect of Financial Index (FI), the index used in the present invention is an index for expressing the operation performance of an enterprise, and may be a repayment ability index, an operation ability index, a profit ability index, a development ability index, a stock price index, and the like. The profitability index may be net asset profitability ROE (Rate of Return on CommonShare's Equisity), asset profitability ROA (Rate of Return on assets), income Per share EPS (earnings Per Share), Market-to-Book Rate MTB (Market-to-Book Rate), and the like.

In the above step 120, the unit of the time interval (T) may be week, month, season, half year, or the like. The time period (N) and time interval (T) are set to collect enough sample data for modeling and verification. If the number of time periods (N) is set to 5, the time interval (T) is set to year, the data collection period (121) is 5 years, and then sample data of 5 years needs to be collected; if the number of time periods (N) is set to 6, the time interval (T) is set to month, the data collection period (121) is 6 months, and then, sample data of 6 months needs to be collected. Because the invention proposes a prediction model, namely the data of the current period is predicted by the data of the previous period, or the data of the next period is predicted by the data of the current period, the time period (N) is at least 2 periods to verify the significance of the prediction result.

In the step 130: if the data collection period (121) is 5 years from 2008 to 2012, the time interval (T) is one year, and the time period number (N) is 5, then the patent index data (131) of all the Patent Indexes (PI) and the financial index data (132) of the Financial Index (FI) of each Patent Entity (PE) in 5 years from 2008 to 2012 must be collected.

In step 140, the panel data, also called parallel data or ensemble data, is a mixture of time series data and cross-sectional data, which refers to data sets of M cross-sectional objects in time period (N), and a total of M × N data sets. Taking this embodiment as an example, if there are 375 Patent Entities (PE), the data collection period (121) is 5 years, i.e. 2008 to 2012, and the time interval (T) is year, then 375 cross-section observed objects are formed, and a data set of 45 Patent Indicators (PI) and Financial Indicators (FI) in 5 time years (2008, 2009, 2010, 2011, 2012) is formed.

Conventional time series data is used to analyze the correlation between observed values (independent variable and dependent variable) of a single observed object at a plurality of times. The conventional cross-sectional data is used to analyze the relationship between observed values (independent variable and dependent variable) of a plurality of observed objects at a single time point. Both types of data are not suitable for the method proposed by the present invention, because the present invention has a plurality of observed objects and a plurality of time points, each of which has a plurality of independent variables and dependent variables. The panel data is used for analyzing the correlation of the observed values of a plurality of observed objects with specific cross sections at a plurality of time points, and the sampling precision of the estimator can be increased, more consistent estimators and effective estimators can be obtained, and more dynamic information can be obtained due to the increase of the observed values, so the panel data is adopted for analysis.

In this embodiment, between the step 140 and the step 150, a normal distribution (NormalDistribution) checking program (145) may be further included to check the normal distribution status of the collected patent index data (131) and financial index data (132). If the data does not show normal distribution, the model is broken down because the error is too high and the data does not converge in the process of establishing the analysis model. Therefore, data that does not exhibit a normal distribution must be subjected to an appropriate conversion operation, converted into a normal distribution state, and then analyzed for independent variables and dependent variables.

The normal distribution test program (145) is commonly used for the following: the Anderson-Darling test program, the Ryan-Joiner test program, the Kolmogorov-Smirnov test program, etc., or the normal distribution of the data can be deduced by simply observing the skewness coefficient and the kurtosis coefficient. For the data not conforming to the normal distribution, a processing operation is required to generate the data of the normal distribution, wherein the Box-Cox conversion procedure is a common method, and the invention is not limited thereto. It must be emphasized that, provided that the raw data already substantially exhibit a normal distribution, no further processing operations need to be applied.

As further illustrated in step 150 above, the distribution curve of the normal distribution is basically a curve with a center line that is symmetrical about the expected value (mean value) and extends to both sides in units of standard deviation. When the mean value and the standard deviation of the two data sets are different greatly, even if the two data sets are distributed normally, the size difference is very large, and analysis and comparison are not facilitated. Therefore, during modeling, the normally distributed data is preferably processed again by a transformation algorithm (152) to normalize the data. Wherein the more commonly used transformation algorithm (152) is a "Z-score" algorithm. The data set calculated by the "Z-score" had an expected value (average value) of 0 and a standard deviation of 1. If all the independent variable and dependent variable data sets are converted into normal distribution and standard score data sets, the relationship is easier to be mined.

In step 150, the first panel data (141) of this embodiment is converted into the second panel data (151) which is normally distributed and standardized and divided after passing through the conversion operation program (152).

In step 160, the first time series calculation program (161) is preferably a one-dimensional Granger causal Test Model. The glange causal test model was pioneered by the nobel economics awarder clevelenge (global w.j.granger) in 2003 for analyzing leading and lagging relationships among time-series economic variables. The basic concept is that if there are two variables X and Y, the variable X occurs first, the variable Y occurs later, and after passing through the glange causal test model, it is verified that the variable X has a significant influence on the occurrence probability of the variable Y, and the variable X is said to be ahead of the variable Y, or the variable X is referred to as a leading indicator of the variable Y. In the economic variables processed by the grangeje causal test model, the independent variable and the dependent variable are all time-series variables, the basic operation model is a regression analysis model, but before the regression analysis, a time offset is set for the independent variable or the dependent variable, namely a lead period or a fall period is set, and then the degree of matching of the regression analysis model with the lead period or the fall period is examined to verify the lead effect or the lag effect of the independent variable.

The model of the Glange causal test model is matched with P _ value obtained by a moderate common F test, and the P _ value is acceptable for the model and is less than 0.1 generally; if P _ value is less than 0.05, the model is good; if P _ value <0.005, the model is excellent. In brief, if the independent variable is X, the dependent variable is Y, and P _ value <0.05 obtained after the grand cause and effect test model indicates that the independent variable X has a leading effect on the dependent variable being Y within a confidence interval of ninety-five percent. Another commonly used model of the glange causal test model is the R-square value, which is between 0 and 1, the better the R-square value is, the worse the R-square value is, the closer to 1, the worse the R-square value is, the closer to 0; 1 is taken as the best, the model is represented to be perfect, and no error exists; with 0 as the worst, the error is infinite.

The number of independent variables represented in the grand causal test model is not limited. That is, the glange causal test model may analyze the lead effect of an independent variable on a dependent variable, which is referred to as a unary glange causal test model, as in step 160; the lead effect of a plurality of independent variables on one dependent variable can also be analyzed simultaneously, and the model is called a multivariate Glanduger causal test model. However, it should be understood that if there is severe co-linearity among multiple independent variables or the data dispersion of the independent variables is too high, the multivariate granger causal test model is often broken down during operation. Therefore, in step 160, preferably, the individual Patent Index (PI) with significant lead effect is mined from 45 Patent Indexes (PI) in step 170 by first applying a unary grand causal test model operation to the financial index data (132) of the individual Patent Index (PI), which is called a patent lead index (172), and excluding other Patent Indexes (PI) with no significant lead effect, so that each of the mined patent lead indexes (172) has significant lead effect on the Financial Index (FI).

In step 170, the setting of the first threshold (171) is critical, and the setting of the first threshold (171) is too strict, so that a patent leading index (172) with significance cannot be mined; the first threshold value (171) is set too loosely, and the Patent Index (PI) which is possibly too much and has insufficient significance is also mistakenly considered as the patent leading index (172) and is mined. If the first time series algorithm (161) takes a univariate glottish causal test model, as described above, the model matches the P _ value obtained from the conventional F test, typically P _ value <0.1, which is acceptable by the model, to a 90% confidence interval; if P _ value is less than 0.05, the model is good and a 95% confidence interval is reached; if P _ value <0.005, the model is excellent, reaching a 99.5% confidence interval. At this time, the first threshold (171) can be set to 0.1, the confidence intervals of how many patent lead indicators (172) reach 90% are initially found out by means of solution, and if the number is not large, the first threshold (171) is set to 0.1; if the number is large, the first threshold (171) can be set to be 0.05 or 0.005, and then a key patent lead index (172) with excellent significance of the lead effect can be found.

In the case that the first threshold (171) is set at 0.05, after the patent lead indicator (172) is mined, if the time lead (L) set in step 160 is one year, the value of the patent lead indicator (172) of each Patent Entity (PE) in the current year is observed, which can be used as the prediction result of the financial index value of the Patent Entity (PE) in the next year and can reach a 95% confidence interval. If the first threshold (171) is set at 0.005 and the time lead (L) set in step 160 is three years, after the patent lead indicators (172) are mined, the value of the patent lead indicator (172) of each Patent Entity (PE) in the current year is observed, which can be used as the prediction result of the financial index data of the Patent Entity (PE) in three years and can reach a 99.5% confidence interval.

The above listed company of the sea exchange is a Patent Entity (PE), and the implementation of the first preferred embodiment will be further described below.

Shanghai exchange, ending in 2012, has 951 listed companies, and Patent Entities (PE) have 951 parents, and the distribution and proportion of industries are shown in FIG. 2.

The invention establishes a prediction model of the precedence of the Patent Index (PI) to the Financial Index (FI), so that the structure of the subsidiary company of the listed company needs to be considered. If the finance of the subsidiary company is incorporated into the parent company for calculation, the Patent Index (PI) of the subsidiary company must also be incorporated into the parent company for calculation. Therefore, 951 listed companies must investigate their subsidiary structure. After investigation, we found that the number of the subsidiary companies in 951 listed companies was 13.6 on average and 9.0 on the median. From the difference between the average and median of the subsidiaries, it can be seen that most of the subsidiaries owned by the public companies are estimated to be less than 10, but a few of them have a very large number of subsidiaries, thus raising the average. Under investigation, the listed companies with the most subsidiaries had up to 174 subsidiaries.

In terms of Financial Index (FI) of listed companies, we initially picked a net asset profitability ROE (Rate of Return on Common Shareholder's Equisity) as a representative for subsequent analysis. For the Patent Indices (PI), 45 Patent Indices (PI) such as P1 to P45 are used.

In terms of selecting valid samples, it is set that Financial Indexes (FI) in five years from 2008 to 2012 all have data, and at the end of 2012, there are at least 50 total patents (including summary of disclosure, invention authorization, utility model, appearance design, etc.), and 951 listed companies have 375 final products after screening to meet the conditions, that is, 375 samples of Patent Entities (PE), whose industry distribution and occupation ratio are shown in fig. 3.

After 375 Patent Entities (PE) collect data (131) of 45 Patent Indicators (PI) and data (132) of Financial Indicators (FI) from 2008 to 2012, the data (141) of the first panel is composed for subsequent analysis.

Referring to fig. 4, for 375 Patent Entities (PEs) of the present invention, the patent index data (131) of 45 Patent Indexes (PI) and the financial index data (132) of the Financial Index (FI) in 5 time years (2008, 2009, 2010, 2011, 2012) form a part of the content of the first panel data (141).

After the first panel data (141) is generated, it is necessary to analyze the data distribution state of the data, whether the data is independent variable data or dependent variable data, and check whether the data conforms to normal distribution. Only data conforming to normal distribution is easy to establish a relational model of independent variables and dependent variables, otherwise, the model is likely to crash due to too large error.

By the Kolmogorov-Smirnov test program, we found that in the first panel data (141), the financial index data (132) of the Financial Index (FI) substantially shows a normal distribution, and the patent index data (131) of each Patent Index (PI) does not show a normal distribution. Therefore, the Box-Cox conversion program is applied to the patent index data (131) of each Patent Index (PI) to make the patent index data approximately show normal distribution.

Then, a Z-score conversion operation program (152) is applied to all the normally distributed Patent Index (PI) patent index data (131) and financial index data (132) of the Financial Index (FI) so that the expected values (average values) of the data are all 0 and the standard deviations are all 1. At this time, the first panel data (141) is converted into the second panel data (151) which is normally distributed and standardized.

In the second panel data (151), the independent variable is 45 Patent Indexes (PI) of 375 Patent Entities (PE) in 5 years, and the dependent variable is a Financial Index (FI) of 375 Patent Entities (PE) in 5 years: ROE, then we used a univariate grand causal Test Model (Granger Causality Test Model) to Test each independent variable against the dependent variable in turn.

One parameter of importance in the Granger causal Test Model (Granger Causality Test Model) is the setting of the time lead period (L). In this embodiment, the time lead period (L) is set to 1 year, 2 years, 3 years, and 4 years, respectively, to observe the lead relationship between the Patent Index (PI) and the Financial Index (FI).

When the set time lead period (L) is 1 year, the used independent variables are Patent Indexes (PI) of four years, namely 2008, 2009, 2010 and 2011; the corresponding dependent variable is Financial Index (FI) of four years 2009, 2010, 2011, 2012. The Patent Index (PI) in 2008 is a Financial Index (FI) paired to 2009, the Patent Index (PI) in 2009 is a Financial Index (FI) paired to 2010, the Patent Index (PI) in 2010 is a Financial Index (FI) paired to 2011, and the Patent Index (PI) in 2011 is a Financial Index (FI) paired to 2012, so that the significance of the advanced effect of the Patent Index (PI) on the Financial Index (FI) can be verified. In this case, 375 × 4 pieces of second panel data (151) can be used, which is 1500 pieces of data.

When the set time lead period (L) is 2 years, the used independent variable is a Patent Index (PI) of three years, namely 2008, 2009 and 2010; the corresponding dependent variables are Financial Indexes (FI) of three years, namely 2010, 2011 and 2012, and in this case, 375 × 3 data is available in the second panel data (151) as 1125 data.

When the set time lead period (L) is 3 years, the used independent variable is a Patent Index (PI) of 2008 and 2009; the corresponding dependent variables are Financial Indexes (FI) of 2011 and 2012 respectively, and in this case, 375 × 2 ═ 750 data in the second panel data (151) can be used.

When the set time lead period (L) is 4 years, the used independent variable is a Patent Index (PI) of 2008; the corresponding dependent variable is Financial Index (FI) of 2012, and only 375 data in the second panel data (151) can be used at this time.

The first time series of operations (161) we use the univariate grand causal Test Model (granger causality Test Model), the Model fitting degree we use the F-Test, the Model fitting degree P _ value uses three first thresholds (171):

(1) p _ value is less than 0.1, and the leading relation of independent variable Patent Index (PI) to dependent variable Financial Index (FI) has acceptable significance and reaches a 90% confidence interval;

(2) p _ value is less than 0.05, and the leading relation of independent variable Patent Index (PI) to dependent variable Financial Index (FI) has good significance and reaches 95% confidence interval;

(3) p _ value is less than 0.005, and the leading relation of the independent variable Patent Index (PI) to the dependent variable Financial Index (FI) has excellent significance and reaches a confidence interval of 99.5 percent.

After the second panel data (151) is operated by a unary grand causal test model, the leadingness of some Patent Indexes (PI) to the Financial Index (FI) is found to be significant, and the second panel data is called as a patent leadingness index (172). Fig. 5 to 8 show the patent lead indicators (172) with significant lead periods of one year, two years, three years and four years, respectively, where P-value < 0.1; represents P-value < 0.05; represents P-value < 0.005.

5-8, if we pay more attention to the investment performance after one year, then observe the value change of the patent lead index (172) with the lead period of one year as shown in FIG. 5; if we pay attention to the investment performance after two years, the numerical value change of the patent leading index (172) with the leading period of two years shown in FIG. 6 is observed; when attention is paid to the investment performance for three or four years, the numerical change of the patent lead indicator (172) shown in fig. 7 or 8 may be observed.

If we are more concerned about the most core patent lead indicator (172) that leads one, two, three, or four years, we can find the patent lead indicator (172) that appears in fig. 5-8 at the same time. The leading indicators (172) of the following 5 patents are significant in the first year, the second year, the third year and the fourth year.

P12 (number of current utility model)

P23 (the total number of pages of the current utility model patent specification)

P29 (claim total of the current utility model)

P35 (the current time utility model patent number)

P41 (number of the current practical patent specification)

The five leading indicators (172) of the patents P12, P23, P29, P35 and P41 all belong to the new type of practical Patent Indicators (PI), because the marketing companies of the shanghai exchanges have the most manufacturing industry (industry code C), and the effective sample 375 Patent Entities (PE) included in the embodiment have the most manufacturing industry, and the patents of the manufacturing industry have most occupied the utility model in the past, so the new type of practical Patent Indicators (PI) naturally and easily stand out in the process of building the prediction model. This result also reveals another message: even though many experts consider that the degree of innovation of the utility model is low and the value is not high, the utility model is a prediction tool for predicting the performance of the Financial Index (FI) of the company on sale.

It must be emphasized here that when the valid samples of the modeling, i.e. the Patent Entities (PEs), change, the last outstanding patent lead indicator (172) changes according to the sample characteristics. For example, when a valid sample of the modeling employs a listed company of the Shenzhen exchange, the standing patent lead indicator (172) must be partially different from the patent lead indicator (172) mined for the sample by the listed company of the above sea exchange. That is, since the patented lead indicator (172) changes according to the sample characteristics, the method of the first preferred embodiment of the present invention is more general and suitable for various sample groups. The information industry marketing company can be used as a sample group to mine a patent leading index (172) suitable for the information industry marketing company; also can be directed against the company that appears in the market of the biological medicine industry as the sample group, excavate the leading index of patent (172) suitable for the company that appears in the market of the biological medicine industry; moreover, the method can independently excavate the patent leading index (172) suitable for the company on which the material chemical industry is listed aiming at the company on which the material chemical industry is listed.

Although the Financial Index (FI) is the equity return rate ROE for the Patent Entity (PE) for the above description and verification of the listed company of the maritime exchange, it should be emphasized that the equity return rate ROE is only for convenience of description, and the method of the first preferred embodiment is applicable to various existing Financial Indexes (FI).

Referring to fig. 9, a second preferred embodiment of the present invention is a method (500) for constructing a patent lead equation (501) for predicting financial information of a Patent Entity (PE). The patent lead equation (501) can generate a patent lead score (502), and the patent lead score (502) leads financial information of the Patent Entity (PE) by a predetermined time lead period (L). The method (500) for constructing the lead equation of this patent comprises the steps of:

step 510: according to the method (100) of constructing the patent lead indicator of the first preferred embodiment, the second panel data (151) and a plurality of patent lead indicators (172) are obtained.

Step 520: third panel data (521) is filtered from the second panel data (151) based on a plurality of patent lead indicators (172).

Step 530: and providing a second time series operation program (531) based on the time lead period (L), wherein the independent variable of the second time series operation program (531) is the lead index (172) of all patents of the third panel data (521), and the dependent variable of the second time series operation program (531) is the Financial Index (FI) of the third panel data (521).

Step 540: setting a second threshold (541), calculating third panel data (521) through a second time series operation program (531) and a time lead period (L), screening a plurality of patent core indexes (542) meeting the second threshold (541) from a plurality of patent lead indexes (172) and generating a patent lead equation (501), wherein the patent lead equation (501) is substantially composed of the plurality of patent core indexes (542) and corresponding weight coefficients (543).

The main purpose of the second preferred embodiment is to combine multiple patent lead indicators (172) into a patent lead equation (501) for the case of multiple patent lead indicators (172), so as to predict the financial performance of the enterprise more quickly and conveniently. A patent lead score (502) is generated by inputting a data value of a patent core index (542) of a Patent Entity (PE) into a patent lead equation (501). The higher the patent lead score (502) is, the higher the value of the corresponding Financial Index (FI) of the Patent Entity (PE) after the time lead period (L) is; the lower the patent lead score (502), the lower the value of the corresponding Financial Index (FI) of the Patent Entity (PE) after the time lead period (L). The Financial Index (FI) of the enterprise directly expresses the operation performance, and the higher the value of the Financial Index (FI), the better the operation performance of the enterprise is, and the more investment value is. Since the patent lead score (502) represents the value of the corresponding Financial Index (FI) of the Patent Entity (PE) after the time lead period (L), observing the high or low of the patent lead score (502) enables to select the objects with investment potential from the Patent Entity (PE).

In step 530, the second time series algorithm (531) is a multivariate grand causal test model with independent variables of the plurality of patent lead indicators (172) obtained in step 170 of the first preferred embodiment, for combining the plurality of patent lead indicators (172) together to discover the lead effect of the combination of the plurality of patent lead indicators (172) on the Financial Indicator (FI).

However, at this time, we must understand another important concept that the multivariate grand causal test model is not a simple summation of the results generated by the multiple univariate grand causal test models, and when combining multiple patent lead indexes (172), the significance of the lead effect of the individual patent lead indexes (172) on the Financial Index (FI) changes, even if the lead effect of some of the patent lead indexes (172) becomes insignificant. Thus, in step 530, the dependent variable may be further manipulated to delete the program item by item, as preferred. That is, all patent lead indicators (172) are first incorporated into the dependent variable of the multiple grand causal test model, the P _ value of each patent lead indicator (172) after the test is observed, the patent lead indicator (172) with the worst significance or even no significance is deleted, then the multiple grand causal test model is redone, the P _ value of each patent lead indicator (172) after the test is observed, the patent lead indicator (172) with the worst significance or even no significance is deleted, and the process is repeated, finally the patent lead indicator (172) with higher significance is left and is called as the patent core indicator (542). At this time, the multivariate grand causal test model integrates all the patent core indicators (542) to generate the patent lead equation (501), and the patent lead equation (501) is substantially composed of a plurality of patent core indicators (542) and corresponding weight coefficients (543).

We will take the patent lead indicators (172) shown in fig. 5, which are the first year old, as an example, and these patent lead indicators (172) are used to predict the financial performance of the Patent Entity (PE). Some of these patent lead indicators (172) can also be used to evaluate a single patent as an indicator of the strength of the single patent, such as:

p6: mean patent life of invention disclosure

P7: mean patent life of utility model patent

P9: mean patent life of patent granted by invention

P10: mean period of examination of patent granted by invention

P15: the average examination period of the patent granted by the invention at the current date is from the application date to the publication granting date

P38: mean number of exclusive rights of current utility model patent

If we want to make the patent lead indicators (172) of P6, P7, P9, P10, P15, P38, etc. leading by one year into the patent lead equation (501), first, we will make the Financial Indicators (FI) of 375 Patent Entities (PE) from 2008 to 2012 in the second panel data (151): ROE and 6 patent lead indicators (172): and P6, P7, P9, P10, P15 and P38 are selected to form third panel data (521).

Next we set a second threshold (541) using a multivariate glange causal test model: p _ value <0.05, the first analyzed independent variable is the 6 patent lead indicator (172): p6, P7, P9, P10, P15 and P38, the dependent variable is ROE, and the analysis result is shown in FIG. 10A. It can be seen that the lead effect of each patent lead indicator (172) changes significantly, and P15 becomes the worst, with P _ value 0.7833.

In the second analysis, we eliminate the worst P15 with P _ value, and the independent variables use 5 patent leading indicators (172) such as P6, P7, P9, P10, and P38, and the analysis result is as shown in fig. 10B. Where P9 is the worst and P _ value 0.4188.

In the third analysis, the worst P9 with P _ value is removed, the independent variables use 4 patent lead indexes (172) such as P6, P7, P10 and P38, the analysis result is shown in FIG. 10C, each patent lead index (172) meets the second threshold (541), the P _ value is less than 0.05, and the prediction model reaches a 95% confidence interval. The 4 patent lead indicators (172) shown in FIG. 10C, at which time we define the patent core indicator (542).

The basic operation model of the multivariate granger causal test model is a multivariate regression analysis model, so the third analysis generates a combined equation besides the excavation of the patent core index (542), which is called as a patent lead equation (501), wherein

Patent lead equation (501) ═ w6 × P6+ w7 × P7+ w10 × P10+ w38 × P38

Wherein w6, w7, w10 and w38 correspond to the patent core indicators (542): the weighting coefficients (543) of P6, P7, P10, P38 represent the sensitivity of their respective patent core indicators (542) to a leading prediction of the Financial Indicator (FI). In this embodiment, it is a good model that w6 is 0.1236, w7 is 0.0236, w10 is 0.0596, and w38 is 0.0247, when the R-square value of the patent lead equation (501) reaches 0.9065, and the adjusted R-square value reaches 0.8750. In this example, the value of w6 is the highest, almost 5 times that of w38, indicating that P6 (mean patent life of the patent disclosure) is most sensitive to the prediction of Financial Index (FI). When the values of P6 (mean patent life of patent publication) and P38 (mean number of independent rights of current utility model patent) are changed by only one unit, the amount of change of P6 (mean patent life of patent publication) to the Financial Index (FI) is 5 times the amount of change of P38 (mean number of independent rights of current utility model patent publication) to the Financial Index (FI).

Referring to fig. 11, a third preferred embodiment of the present invention is a method (600) for evaluating investment potential of an enterprise, comprising the following steps:

step 610: collecting patent information (612) of a plurality of enterprises (611);

step 620: providing a patent lead equation (501) obtained by the second preferred embodiment, wherein the patent lead equation (501) is substantially composed of a plurality of patent core indicators (542) and their corresponding weighting coefficients (543);

step 630: calculating data (631) of a patent core index (542) corresponding to each enterprise based on the patent information (612) of each enterprise (611);

step 640: generating a patent lead score (502) for each enterprise (611) by calculation of a patent lead equation (501) based on data (631) of a patent core index (542) for each enterprise (611);

step 650: the patent lead scores (502) of the enterprises (611) are ranked by a ranking program (651), and the ranking result (652) represents the investment potential of the enterprises (611).

The higher the lead score (502) of the above patent, the higher the value of the corresponding Financial Index (FI) of the enterprise (611) after the time lead period (L); the lower the patent lead score (502), the lower the value of the corresponding Financial Index (FI) of the enterprise (611) after the time lead period (L). The Financial Index (FI) of the enterprise (611) has a higher value, so that the higher the value of the Financial Index (FI), the better the operation performance of the enterprise is, and the higher the investment value is; the lower the value of the Financial Index (FI), the worse the enterprise operation performance, and the less investment value. Since the patent lead score (502) represents the value of the corresponding Financial Index (FI) of the enterprise (611) after the time lead period (L), the enterprise (611) can select the objects with investment potential according to the high and low ranks of the patent lead score (502).

Referring to fig. 12, a fourth preferred embodiment of the present invention is a computer system (700) for evaluating investment potential of an enterprise, which includes a patent information collecting device (710), an index calculating device (720), a patent lead score calculating device (730), and a score ranking device (740).

The patent information acquisition device (710) collects patent information (612) of a plurality of enterprises.

An index calculation device (720) calculates and generates data (631) of a patent core index (542) corresponding to each enterprise based on patent information (612) of each enterprise.

The patent lead score calculating means (730) calculates the patent lead score (502) of each enterprise according to the patent lead equation (501) obtained in the second preferred embodiment according to the data (631) of the patent core index (542) of each enterprise, wherein the patent lead equation (501) substantially consists of a plurality of patent core indexes (542) and corresponding weight coefficients (543).

The score ranking device (740) ranks the plurality of patent lead scores (502), and the ranking result (652) represents the ranking of the investment potential of the plurality of enterprises.

The method (600) and the computer system (700) for evaluating the investment potential of the enterprise are based on big data, objective operation and rigorous verification, are not only beneficial to the development of the technical strength of patent information analysis and utilization, but also can promote the positive development of the investment method in the investment field, and have positive supporting effect on the research, development and innovation of industrial technology.

The foregoing description should be understood and implemented by those skilled in the relevant art. Meanwhile, the above description is only a preferred embodiment of the present invention, and is not intended to limit the scope of the present invention. Any equivalent changes or modifications made based on the disclosure of the present invention should be included in the scope of the claims.

Claims (10)

1. A method of assessing investment potential of an enterprise, comprising:

(a) collecting patent information of a plurality of enterprises;

(b) providing a patent lead equation, which consists of a plurality of patent core indexes and corresponding weight coefficients thereof, wherein the patent lead equation is used for predicting financial information of a patent entity, the patent lead equation generates a patent lead score, the patent lead score leads the financial information of the patent entity by a preset time lead period, and the patent lead equation is obtained by a construction method of the patent lead equation;

(c) calculating data of corresponding patent core indexes of each enterprise based on the patent information of each enterprise;

(d) generating a patent lead score of each enterprise through calculation of the patent lead equation based on the data of the patent core indexes of each enterprise; and

(e) ranking the plurality of patent lead scores through a ranking program, wherein the ranking result represents the investment potential of the plurality of enterprises;

the patent lead equation constructing method includes the following steps:

(1) setting a plurality of patent entities and a plurality of patent indexes and financial indexes for describing each patent entity, wherein each patent index is obtained by the operation of the patent information of each patent entity, the patent indexes comprise indexes which are calculated by total number and current time and are used for describing Chinese invention disclosed patents, indexes for describing Chinese invention authorized patents, indexes for describing Chinese utility model patents and indexes for describing Chinese appearance design patents, the financial indexes are stock price indexes, and the current time calculation means that the data operation of the patent indexes is limited in a certain specific time interval;

(2) setting a data collection period, wherein the data collection period consists of a time interval and a time period number, and the time period number is an integer not less than two;

(3) collecting the corresponding patent index values and financial index values of each time interval of each patent entity in the data collection period;

(4) forming a first panel data by the plurality of patent index values and the plurality of financial index values of the plurality of patent entities;

(5) forming second normally distributed and standard fractional panel data by converting the first panel data, wherein the conversion is a Z-fraction operation;

(6) setting the time-lead period and providing a first time-series calculation procedure based on the time-lead period, the time-lead period including at least one of the time intervals, the independent variable of the first time-series calculation procedure being one of the patent indicators of the second panel data, the dependent variable of the first time-series calculation procedure being the financial indicator of the second panel data, wherein the first time-series calculation procedure is a unary grand causal test model;

(7) setting a first threshold, using the first time sequence operation program and the time lead period successively, operating the second panel data, and screening at least one patent lead index meeting the first threshold from the plurality of patent indexes, wherein the first threshold is not more than 0.1;

(8) screening and forming third panel data from the second panel data based on the plurality of patent lead indexes;

(9) providing a second time series operation program based on the time lead period, wherein the independent variable of the second time series operation program is all the patent lead indexes of the third panel data, and the dependent variable of the second time series operation program is the financial index of the third panel data, wherein the second time series operation program is a multivariate grand cause and effect inspection model;

(10) setting a second threshold, calculating the third panel data through the second time series calculation program and the time lead period, screening a plurality of patent core indexes meeting the second threshold from the plurality of patent lead indexes, and generating a patent lead equation, wherein the patent lead equation is composed of the plurality of patent core indexes and corresponding weight coefficients.

2. The method for assessing investment potential of a business as claimed in claim 1, wherein the plurality of patent entities are listed companies.

3. The method for assessing investment potential in a business as claimed in claim 1, wherein the time interval is selected from the group consisting of: week, month, season, half year, and year.

4. The method for assessing investment potential in an enterprise as claimed in claim 1, wherein between steps (4) and (5), further comprising a normal distribution test procedure for testing any of said patent index values.

5. The method for assessing investment potential in a business as claimed in claim 1, wherein between steps (4) and (5), further comprising a normal distribution test procedure for testing the value of the financial index.

6. A computer system for evaluating enterprise investment potential comprises a patent information acquisition device, an index calculation device, a patent lead score calculation device and a score sorting device, and is characterized in that: the patent information acquisition device collects patent information of a plurality of enterprises; the index calculation device calculates and generates data of the corresponding patent core indexes of the enterprise based on the patent information of each enterprise;

the patent lead score calculating device generates the patent lead score of each enterprise through calculation of a patent lead equation according to the data of the patent core indexes of each enterprise, wherein the patent lead equation consists of a plurality of patent core indexes and corresponding weight coefficients, the patent lead equation is used for predicting financial information of a patent entity, the patent lead score leads the financial information of the patent entity by a preset time lead period, and the patent lead equation is obtained by a construction method of the patent lead equation; the score sorting device sorts the patent lead scores to generate a sorting result, and the sorting result represents the investment potential of the enterprises;

the construction method of the patent lead equation comprises the following steps:

(1) setting a plurality of patent entities and a plurality of patent indexes and financial indexes for describing each patent entity, wherein each patent index is obtained by the operation of the patent information of each patent entity, the patent indexes comprise indexes which are calculated by total number and current time and are used for describing Chinese invention disclosed patents, indexes for describing Chinese invention authorized patents, indexes for describing Chinese utility model patents and indexes for describing Chinese appearance design patents, the financial indexes are stock price indexes, and the current time calculation means that the data operation of the patent indexes is limited in a certain specific time interval;

(2) setting a data collection period, wherein the data collection period consists of a time interval and a time period number, and the time period number is an integer not less than two;

(3) collecting the corresponding patent index values and financial index values of each time interval of each patent entity in the data collection period;

(4) forming a first panel data by the plurality of patent index values and the plurality of financial index values of the plurality of patent entities;

(5) forming second normally distributed and standard fractional panel data by converting the first panel data, wherein the conversion is a Z-fraction operation;

(6) setting the time-lead period and providing a first time-series calculation procedure based on the time-lead period, the time-lead period including at least one of the time intervals, the independent variable of the first time-series calculation procedure being one of the patent indicators of the second panel data, the dependent variable of the first time-series calculation procedure being the financial indicator of the second panel data, wherein the first time-series calculation procedure is a unary grand causal test model;

(7) setting a first threshold, using the first time sequence operation program and the time lead period successively, operating the second panel data, and screening at least one patent lead index meeting the first threshold from the plurality of patent indexes, wherein the first threshold is not more than 0.1;

(8) screening and forming third panel data from the second panel data based on the plurality of patent lead indexes;

(9) providing a second time series operation program based on the time lead period, wherein the independent variable of the second time series operation program is all the patent lead indexes of the third panel data, and the dependent variable of the second time series operation program is the financial index of the third panel data, wherein the second time series operation program is a multivariate grand cause and effect inspection model;

(10) setting a second threshold, calculating the third panel data through the second time series calculation program and the time lead period, screening a plurality of patent core indexes meeting the second threshold from the plurality of patent lead indexes, and generating a patent lead equation, wherein the patent lead equation is composed of the plurality of patent core indexes and corresponding weight coefficients.

7. The computer system for assessing investment potential of a business as recited in claim 6, wherein the plurality of patent entities are listed companies.

8. The computer system for assessing investment potential of a business as recited in claim 6, wherein the time interval is selected from the group consisting of: week, month, season, half year, and year.

9. The computer system for assessing investment potential in an enterprise as claimed in claim 6, wherein between steps (4) and (5), further comprising a normal distribution test program for testing any of said patent index values.

10. The computer system for assessing investment potential in an enterprise of claim 6, wherein between steps (4) and (5), further comprising a normal distribution test procedure for testing the value of the financial index.

Images