Develop NLP tool and ifrastructure for extracting hyponymes

The purpose of the task is development the tool for extracting hyponymes for an input word collocations based on determined list of word collocations. An interface for communication with the program will be performed via the web-service which return the .josn encoded list of huponymes for incoming GET request. Requirements: We hope the tool can achieve 80% accuracy rate and 40% recall tate. The freelancer should complete the following milestones. Milestones: 1. 1.5 weeks Finish NLP tool development. Demonstration of the results on a test sample and on collocations from the company's dataset. 2. 0.5 week Taxonomization of an input dataset. Demonstration of the semantic patterns inside the input dataset. 3. 2 weeks Development data infrastructure. Demonstration API for extracting word's hyponymes from the input dataset and API for streaming update of the taxonomized internal dataset. Please update us about the project status every 3 days. You will be responsible and liable for all delays. If you want to extend the time, please inform me in advance.