Machine learning-based patent quality metric
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.
1. A machine-learning based artificial intelligence device for use in finding an estimate of patent quality, the device comprising:
a patent data retriever configured to receive a first set of patent data, the first set of patent data comprising at least one of patent application data and patent data for a plurality of patents, and to generate a list of binary classifiers based on the first set of patent data;
a quantitative data scalar configured to assign a standardized scaled score to each binary classifier of the list of binary classifiers;
a binary classifier optimizer configured to generate, using an automated processor, a candidate set of binary classifiers from the list of binary classifiers using a heuristic search that maximizes a yield according to a cost function in an iterative process to provide a final set of binary classifiers, wherein
the final set of binary classifiers is validated using a second set of patent data comprising at least one of patent application data and patent data for a second plurality of patents; and
a user information manager configured to receive patent information for a target patent and to provide the estimate of patent quality associated with the target patent according to the final set of binary classifiers.
2. The device of claim 1 , wherein the heuristic search comprises an artificial neural network model.
3. The device of claim 2 , wherein maximizing the yield includes iteratively changing a number of hidden layers of the artificial neural network model to maximize area under the curve.
4. The device of claim 1 , wherein maximizing the yield includes iteratively using a genetic algorithm.
5. The device of claim 1 , wherein maximizing the yield includes iteratively using an artificial neural network model and a genetic algorithm.
6. The device of claim 1 , wherein the cost function is a receiver operating characteristic and the yield is calculated according to an area under a curve.
7. The device of claim 1 , wherein the estimate of patent quality represents an estimate of a lifetime of the patent.
8. The device of claim 1 , wherein the patent data retriever is configured to receive the second set of patent data comprising at least one of patent application data and patent data for the second plurality of patents.
9. A system comprising the device of claim 1 and a second device communicatively connected to the machine-learning based artificial intelligence device over a network, the second device comprising:
a second automated processor;
a user interface receiving the target patent information for the target patent;
an estimate requester requesting from the machine-learning based artificial intelligence device, the estimate of patent quality for the target patent; and
a user interface providing, to a user, information indicating the estimate of patent quality.