[Two] canteen project selection team

Project Introduction

• Project name: Text of entities and relationships online tagging system
• Project Description:
  • Text extracted achieve expertise knowledge of the system for a specific area on the internet, for a given specialized text books on specific terminology (entity) in the text refers to various terms and relationships between objects denoted by substituting construct specialized knowledge map of the field, but also can be used as training data to use machine learning.
  • Support multi-user collaborative annotation, annotation data can be recorded and each source synchronization update; solid support for text labels corresponding link, to unfold through a graphical interface on the annotation result, two-way support and positioning.

NABCD

• Need

  For students, in reading some professional books often encounter such a problem: always remember some of the concepts of terms, when confronted with these terms always have to turn back to the previous section view; or troublesome, direct dig the phone Baidu related terms. This time we need a knowledge map to help us understand these concepts, but for starters, is not to build knowledge map is a little difficult?

  When the end of the review, we usually do: go over textbook content and a handwritten mind map to consolidate memories. But the handwritten mind map also has great limitations: difficult to modify, in order to devise a clear and beautiful patterns of knowledge tend to cost us a lot of energy.

  The teacher will encounter such a problem in teaching: Want to show relevant knowledge structure on the PPT, but it is very relevant professional knowledge, one by one to draw it up in a lot of trouble in the mind map.

  To sum up: the mapping knowledge too hard to do!

  So, we need an application that can help us easily generate knowledge maps.

• Approach

  • Djano using Python framework to achieve the main application
  • Echart using JavaScript charting library marked the completion of a graphical display of results

• Benefit

  Project in function can provide users with the convenience of:

  • Click to view the terms of direct support in the text interpretation, providing a smooth reading experience.
  • The term may be directly labeled in the text, a graphical interface and display the results marked. A key to generate your knowledge maps, worry and effort.
  • Support multi-user collaborative tagging, when faced with large-scale text, can effectively improve work efficiency.

  Compared to the advantage of the knowledge map and office handwritten mind map:

  • Text only need to import the application easy to use.

  • The knowledge generated maps can drag yourself to design a look that you want.

• Competitors

  • The Annotator-chinese ( the Git address )

    Introducing the project, this project uses NLP to achieve a smart label active learning algorithm, enables automatic annotation of text. In 2017 the project has been approved, and still do not have a complete result.

  • IEPY ( Git address )

    

    IEPY is also an active learning information extraction and relation extraction tool. The whole project is more complete, user management systems. The front end a little heavy, not very user-friendly.

  • BRAT ( Git address )

    brat is a web-based tool for text annotation; that is, for adding notes to existing text documents.

    

    

    brat specifically designed for structural annotation, wherein the annotation is not free-form text, but has a fixed format may be automatically processed by a computer and interpreted.

  At present, the text annotation tool is mostly based on active learning NLP tools that are functionally more toward services for the machine-learning text labels. We also found no purpose for the establishment of the knowledge map text annotation tool, at this point, our project has great advantages.

• Delivery

  • No public official by major colleges and universities recommend, show students the knowledge map construction of a new application.
  • Publicity CSDN, CNblog technology forum.

Users estimates

• Where the amount of user software release, a week after the release of estimates (accurate to one).

  Due to the nature of the online multi-user operation, we intend to make a website, online publication.

  Since the target audience is basically a college student, taking into account the basic project Published anastomosis with the test, the test of high school students may not take the effort to try new software, so the first week of users optimistic estimate of 100 people.