Machine learning-based patent quality metric

Abstract

A machine-learning based artificial intelligence device for finding an estimate of patent quality, such as patent lifetime or term is disclosed. Such a device may receive a first set of patent data and generate a list of binary classifiers. A candidate set of binary classifiers may be selected and using a heuristic search, for example an artificial neural network (ANN), a genetic algorithm, a final set of binary classifiers is found by maximizing iteratively a yield according to a cost function, such an area under a curve (AUC) of a receiver operating characteristic (ROC). The device may then receive patent information for a target patent and report an estimate of patent quality according to the final set of binary classifiers.

Description

CROSS-REFERENCE TO RELATED APPLICATION[0001]The present non-provisional patent application claims the benefit of priority from U.S. Provisional Patent Application No. 61 / 928,806, filed Jan. 17, 2014, the entire contents of which are incorporated herein by reference.BACKGROUND[0002]1. Field of the Invention[0003]The present disclosure relates to a system comprising a CPU, storage and database of patent grants or applications and other relevant data for computation of an estimation of patent quality utilizing machine learning algorithms for factor selection and classification based on non-linear models.[0004]2. Related Art[0005]Attempts have been made to assess or to estimate the value or expected life of a patent or a patent application based on historic data about patents. However, testable and reproducible quantitative metrics are difficult to come by. Also, using a combination of quantitative factors available from a universe of patent information to arrive at a patent value or es...

AI Technical Summary

Benefits of technology

The patent describes a machine learning system that can estimate the quality of a patent based on various factors such as text data and quantitative data. The system uses a combination of artificial intelligence techniques like binary classifiers, quantitative scoring, and optimizations to make the estimate. The system can also test the validity of the estimated quality using a separate set of data. Additionally, the system can provide an estimate of the patent's lifespan to users based on the final set of binary classifiers. Overall, the system offers a technical solution for efficiently and accurately evaluating the quality of patent data.

Problems solved by technology

However, testable and reproducible quantitative metrics are difficult to come by.

Also, using a combination of quantitative factors available from a universe of patent information to arrive at a patent value or estimated patent life or the like is difficult given the sheer number of patent-related and patent application-related factors and given that each patent represents a unique invention.

Therefore, finding the combination of factors that produces an optimal or maximized patent quality / patent life profile has been a difficult task.

Generally, using a brute-force approach, each additional factor, or combination of factors, increases the complexity and the processing time exponentially.

Images