Forecasting Patent Filings at the European Patent Office (EPO) using - - PowerPoint PPT Presentation

Forecasting Patent Filings at the European Patent Office (EPO) using compositional data analysis techniques.

Peter Hingley, Financial Controlling and Statistics, European Patent Office, M unich, Germany

phingley@epo.org

CoDaWork 2017 Workshop Abbadia San Salvatore 8th June 2017

Disclaimer: The forecasts that will be mentioned are not official forecasts of EPO.

Contents

• 1. Patenting at EPO and the forecasting problem.
• 2. A Dynamic Log Linear (DLL) model for Total (patent) Filings (TFs).
• 3. Fitting the DLL model to TFs.
• 4. Fitting a straight line ilr regression model to Industrial Areas (IA)

proportions.

• 5. Fitting a straight line ilr regression model for TFs with IA proportions

as added predictors.

• 6. Conclusions

European Patent Office

Patentability under the European Patent Convention (EPC)

Patents are granted for inventions in all fields of technology To be patentable, inventions must be new, involve an inventive step and be industrially applicable They must relate to a product, process, apparatus or use. Some things are excluded from patentability (E.G. Discoveries; Scientific theories, Mathematical methods, Computer programs, ....)

3

• 1. Patenting at EPO and the forecasting problem.

Patents are an incentive for economic growth.

§ Makes the latest technological knowledge available to the public § Inspires further innovation § Helps to prevent duplication of R&D § Helps identify new partners and allows licensing § Gives patent holders time to recoup their development costs

European Patent Office

1) By claiming priority of an earlier application filed with a national office (or at WIPO) within 12 months.

4

• 1. Patenting at EPO and the forecasting problem.

First filing Subsequent filing within 12 months First filing + 30 months

Total Filings (TFs) are the sums of these areas.

First filings (Euro-PCT-INT) can also be done at EPO but numbers are small enough to ignore.

European Patent Office

There are various ways to classify patents by technical areas. The International Patent Classification (IPC) system is well established.

IPC has 8 main classes (A to H) and many sub-classes. IPC assignment is by technical content. Assigning to industries of applicants is also possible.

5

• 1. Patenting at EPO and the forecasting problem.

European Patent Office

6

CODA approach to relationships between IPC proportions.

Principal components (or regressions) could be useful for technical trends in patent applications. But this does not solve the forecasting problem for T

• tal Filings (TFs) at EPO.
• 1. Patenting at EPO and the forecasting problem.

2–way breakdowns (Blocs vs IAs) are not considered advisable.

• 1. Patenting at EPO and the forecasting problem.

Breaking down TFs into just three Industrial Areas (IAs), Electricals, Chemicals and Traditionals is more manageable for forecasting purposes. The numbers of Filings with no IA assigned increases after 2013.

Country 1 Country 2 Autoregre- Factor Intercept Intercept ssive terms B 1

• COUNTRY 1

A 1

• 1
• B

DIVIDE BY POPULATION 1

• SIZE OF COUNTRY,

1

• TAKE LOGARITHMS

1

• 1
• 1
• COUNTRY 2

A 1

• 1
• B

1

• DESIGN MATRIX X

Time

.

Linear Model for filings A: Y = X . B

.

Estimate B = (XTX)-1 . XTY

.

Fit & Forecast Y = X . B Back transform per country, add filings estimates per year 8

Dynamic log linear models

• 2. A Dynamic Log Linear (DLL) model for EPO Total filings for patents (TFs).

Series A: Filings to be forecast Series B, C ...: Explanatory variables

Country of origin breakdowns are built in.

European Patent Office

9

where P is the number of EPO T

• tal filings from a country;

L is the number of workers in a country;

i is the country; t is time (years); -1 and -2 indicate lags of one year or two years;

R is R&D expenditures - a stock variable with components from 0 to 5 year lags; The GDP of the source country Y is split into two components:- Y

T is the “ trend” level of output

u is the business cycle variable;

α terms are estimable parameters (αi is a country intercept); εit is an error term, assumed to be normal with constant variance;

ln( ) denotes natural logarithm; and ∆ indicates year-to-year differences.

An approach including the Influence of Business Cycles.

• 2. A Dynamic Log Linear (DLL) model for EPO Total filings for patents (TFs).

European Patent Office

10

Include two Isometric Log-Ratio (ilr) terms to a cut-down version of the DLL model.

Fit two ilr terms rather than three to avoid collinearity. Training data run from year 2001 up to year 2015. M odels fitted up to end of calendar year two years before. So in January 2017, the model was fitted up to 2015.

• 2. A Dynamic Log Linear (DLL) model for EPO Total filings for patents (TFs).

• 3. Fitting the DLL model to TFs.

Parameter estimates after fitting the various models.

* indicates approximate significance at the 95 percent level.

Separate Industrial Areas (IAs) Total Filings Parameter estimates Electricity (E) Chemistry (C) Traditional (T) With IAs Without IAs Autoregression α2

• 0.01
• 0.18*
• 0.05
• 0.09*
• 0.08*

R&D Stock α3 0.29 0.27 0.09 0.36* 0.36* GDP trend α4

• 0.25

2.51* 1.63* 1.26* 1.12* GDP cycle α5 0.39 0.2 0.44 0.53 0.46 ilr for Electricals α6 0.30* ilr for Chemicals α7 0.21* Observation standard error 0.1877 0.1653 0.1325 0.111 0.113 Residual degrees of freedom 388 388 388 386 388 M odel with IAs fits better than w/ o IAs. No significant parameters. Significant

α2 & α4

parameters. Significant

α4

parameter.

• 3. Fitting the DLL model to TFs.

Total Filings forecasts, both with and without ilr terms for IAs, are almost exactly the same. (M odel fit is better with ilr terms). These forecasts are lower than cumulating the forecasts by IAs.

• 4. Fitting a straight line ilr regression model to IA proportions.

Forecasts of proportions of IAs in TFs, based on: - Straight line regressions to the raw proportions; Straight line regressions after transforming proportions to ilrs; and Straight line regressions after transforming proportions to ilrs and including TFs as an additional predictor. (Back-transformation of fitted ilrs to proportions was done in Excel.)

European Patent Office

14

Back transformation of predicted isometric logratio (ilrj) terms to proportions (pj).

• 4. Fitting a straight line ilr regression model to IA proportions.

But what is k? Create a column of trial values of k, k1 say, of sufficient accuracy. Calculate Estimate k by minimising | k1 – k2| .

• 5. Fitting a straight line ilr regression model for TFs with IA proportions as added predictors.

European Patent Office

16

• 1. When modelling TFs, the fit of the DLL model is improved by adding terms

based on IAs. Similarly, a straight line regression for TFs gives improved fit by adding terms based on IAs. BUT In both cases, forecasts of TFs are hardly changed by including the IA terms.

• 2. When modelling proportions of IAs by straight line regressions of ilr based

terms, the results are almost the same as fitting straight lines directly to the

• proportions. BUT The CoDa approach will work better with many classes or

low/ high proportions for some classes.

• 3. Possible methods have been identified to add a Total to a model of

proportions (in straight line regressions of ilrs), or ilr proportions to models for the total (in DLL model and straight line regressions). BUT Here, so far, forecasts are hardly changed although model fits improve.

• 6. Conclusions.

European Patent Office

17

• 4. Theoretical modelling including simulations could better resolve issues

raised under point 3. Perhaps an iterative approach can model both Totals and proportions simultaneously.

• 5. A suggestion to use a “Total” that relates to a geometric mean could be

explored using analysis as at point 4. This could further improve the agreement of forecasts with and without additional terms. Any remaining discrepancy could be due to the approximation of modelling a total from the geometric mean.

• 6. Extensions of the CoDa approach to modelling patent filings (Totals and/ or

proportions) could investigate other breakdowns, like first/ subsequent filings and patent families.

• 6. Conclusions.

THANK YOU

European Patent Office

18

• 1. G. Coenders, B. Ferrer-Rosell, G. M ateu-Figueras and V. Pawlosky-Glahn, V., 2015. “M ANOVA of

Compositional Data with a Total”, Proceedings of the 6th international workshop on compositional data analysis, CoDaWork 2015, www.compositionaldata.com/ codawork2015/ images/ ProceedingsBook.pdf

• 2. EPO website and annual reports. www.epo.org/
• 3. EPO Patent Filings surveys. www.epo.org/ service-support/ contact-us/ surveys/ patent-filings.html
• 4. P

. Hingley & M . Nicolas, 2014. “M ethods for forecasting numbers of patent applications at the European Patent Office”, World Patent Information, 26, pp. 191-204.

• 5. P

. Hingley and W. Park, 2015. “A dynamic log-linear regression model to forecast numbers of future filings at the European Patent Office”, World Patent Information, 42, pp. 19-27.

• 6. P

. Hingley & W. Park, 2016. “Do business cycles affect patenting? Evidence from European Patent Office Filings”, Technological Forecasting and Social Change, 116, pp. 76-86.

• 7. IP5, 2016. (EPO, J

PO, KIPO, SIPO, USPTO) “IP5 statistics reports”, www.fiveipoffices.org/ statistics/ statisticsreports.html

• 8. V. Pawlowsky-Glahn, J

. Egozcuea and D. Lovell, 2014. “Tools for compositional data with a total”., Statistical M odelling, 15, pp. 175-190. / See also Proceedings of the 5th international workshop on compositional data analysis, CoDaWork 2013, www.statistik.tuwien.ac.at/ CoDaWork/ CoDaWork2013Proceedings.pdf

• 9. World Intellectual Property Organization, 2017. “International Patent Classification”,

http:/ / www.wipo.int/ classifications/ ipc/ en

References.