Skip to content

Data Labeling System created by Talha Bayburtlu, Can Karatepe, Çağan Oduncuoğlu, Emre Aydemir, Osman Erikçi and Batuhan Karataş within Object Oriented Software Design course.

Notifications You must be signed in to change notification settings

talhabayburtlu/Data-Labeling-System

Repository files navigation

Data Labeling System

Data Labeling System created by Talha Bayburtlu, Can Karatepe, Çağan Oduncuoğlu, Emre Aydemir, Osman Erikçi and Batuhan Karataş by using Scrum model.The project is built within Object Oriented Software Design course in Computer Engineering Department. Also project is checked by Murat Can Ganiz in each iteartion.

Description

This project's main idea is to label data by using some in-built mechanisms. Those mechanism are manual mechanism,random mechanism, keyword based mechanism and auto label mechanism. Their descriptions are:

  • Manual mechanism is for human users which presents the instances one by one and the labels that are available for that dataset.
  • Random mechanism is for bot users which labels instances randomly. Random mechanism is mainly used for first iteration to built the skeleton of the project.
  • Keyword based mechanism is used for bot users which labels instances by looking if given words are available in instance. If it is, the corresponding labels are assigned to that instance.
  • Auto label mechanism is for bot users which is simular to keyword based mechanism. The main difference is there are no keywords to look up, but a search done for each instance and if available labels are contained by that instance, corresponding labels are assigned.

There are also some statistics about mainly for users, datasets and instances that stores information about their relations. The statistics includes information like a user's consistency, the variety of assigned labels for particular instances, the completion and consistency percantages.

Iteration 1

In this iteration the skeleton of the project is mainted by using random labeling mechanism. Entites are created, their relations are made and entites tested by random labeling mechanism.

Iteration 2

This iteraion's main idea is to create statistics about entites and their relationships between them.

Iteratino 3

We created a ui for human users to interact with data labeling mechanism. Also created different types of labeling mechanisms for both human and bot users.

Collabrators

  • 150117038 Osman Erikci
  • 150117066 Zekeriya Batuhan Karataş
  • 150118004 Can Karatepe
  • 150118036 Abdullah Emre Aydemir
  • 150118066 Talha Bayburtlu
  • 150118070 Melik Çağan Oduncuoğlu

About

Data Labeling System created by Talha Bayburtlu, Can Karatepe, Çağan Oduncuoğlu, Emre Aydemir, Osman Erikçi and Batuhan Karataş within Object Oriented Software Design course.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Contributors 6

Languages