We are looking for talents who can use NLP&ML to make a text classification program for us.
Classification standard: There are 22 main classes about different industries according to our classification standard. Under every main class are some subclasses. There about 170 subclasses overall. Every inputted text should be put into at least one of the 170 classes. A text can be put into 2 classes (maximum) if the two classes are both related.
Accuracy: The classification accuracy should be no less than 80%.
Input: The inputted text is commercial description from different industries, including: news, company overview, business for sale, etc.
Output: The name(s) of the class.
Training set: We can provide raw materials about company overview and business for sale. For every class we have only 10 tagged materials with high classification accuracy.
Coding language: Not limited, but Python and Java are preferred.
We will reveal more when contacting with you. Please leave us message or contact so I can discuss with you.