AI Patent Classifier II , How it works with AI

 

In the last issue #13, we introduced WIPS' new product, AI patent classifier.

The classification criteria for classifying patent differ depending on the purpose of the analysis, and we have found that it is very important to obtain a result that meets the purpose. We introduced that the WIPS’ AI ​​patent automatic classifier is a product that generates a classification engine optimized for the user's criteria by learning the user's correct answer data using 5 specialized AI algorithms.

Since the introduction of the classifier was released, we’ve received many inquiries about the product and how it works. This time, more information is open.

 The AI ​​patent classifier developed by WIPS is an application that classifies unclassified patent documents in accordance with the user-defined criteria.

The principle of the product in brief,

①     The combination of 5 algorithms creates ②70 untrained models (INDUCER), and ③ inputs 3 data types and it makes ④ 210 individual classification models.

From the 210 classified models, ⑤ 15 final models are selected based on the measured values, ⑥and the final best model is determined by applying 3 ensemble types.

This is a user-customized classification engine created by training the input data. Here, ‘training (also learning)’ refers to the process of creating a classification model by applying the user-defined correct answer data to the inducer.

The classification engine created through the above process appears on the screen as shown above. Now just upload the data you need to classify here.

Unclassified data is automatically classified with the user's criteria through an inference process in the created classifier.

(*Inference: the process by which the classification model predicts individual categories for unclassified patent documents)

Individual engines created by the inference process can be used continuously when the relevant patents are updated that meet the same purpose and criteria. For example, if there is a classification engine created to analyze the structure of a robot vacuum cleaner, automatic classifying becomes possible within a short period of time when the technology of the suction part or control of the robot cleaner is filed.

You can have the free trial right away.

Just prepare only two files, 1) correct answer data to train AI and, 2) unclassified data to be inferred by the classifier that has trained the correct answer.

 Try it right now!