Crowdsourcing Marketplace for Text Creation Tasks Case Study: Text Summarization Tasks
<p align=”justify”>The case study used in this final project is the text summarization tasks, there has been algorithm that help ease the process, which is using Automatic Text Summarization. However, the algorithm is not able to produce precise and accurate work results and manmade summ...
Saved in:
Main Author: | |
---|---|
Format: | Final Project |
Language: | Indonesia |
Online Access: | https://digilib.itb.ac.id/gdl/view/30856 |
Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
Institution: | Institut Teknologi Bandung |
Language: | Indonesia |
id |
id-itb.:30856 |
---|---|
spelling |
id-itb.:308562018-07-02T10:15:34ZCrowdsourcing Marketplace for Text Creation Tasks Case Study: Text Summarization Tasks Anglila Hapsari - NIM : 13514069 , Sekar Indonesia Final Project INSTITUT TEKNOLOGI BANDUNG https://digilib.itb.ac.id/gdl/view/30856 <p align=”justify”>The case study used in this final project is the text summarization tasks, there has been algorithm that help ease the process, which is using Automatic Text Summarization. However, the algorithm is not able to produce precise and accurate work results and manmade summaries still produces many errors and eliminate the automation aspects that exist in automatic text summarization. The solution to this problem is crowdsourcing, which is outsourcing tasks in the form of open calls online. However, there are no crowdsourcing platforms focused on text creation tasks. Meanwhile, text creation tasks require a workflow that may be different than other tasks in general. This final project examines workflows, task decomposition methods, and quality management appropriate for text creation tasks, using the case study text summarization tasks. <br /> <br /> <br /> <br /> <br /> In this final project, a crowdsourcing marketplace is created using the prototyping method. It must meet all the requirements of crowdsourcing text creation tasks. We tested two kinds of workflow and task decomposition methods, namely Partition-Map-Reduce and Worker Based Task Decomposition, as well as two kinds of quality management, namely Find-Fix-Verify and Evaluation by Experts with Real-time Support. <br /> <br /> <br /> <br /> <br /> This final project implementation resulted a crowdsourcing marketplace for text creation tasks called Ensemble. The results of functional testing show that out of fourteen functional requirements, nine have been fulfilled, while 80% -90% of workers agreed that the remaining five functional requirements have been met. The test results show that the Partition-MapReduce the workflow and task decomposition method, as well as quality management in the form of Evaluation by Expert and real-time support can be used for text creation tasks, and is the best for the case study and sample population in this final project.<p align=”justify”> text |
institution |
Institut Teknologi Bandung |
building |
Institut Teknologi Bandung Library |
continent |
Asia |
country |
Indonesia Indonesia |
content_provider |
Institut Teknologi Bandung |
collection |
Digital ITB |
language |
Indonesia |
description |
<p align=”justify”>The case study used in this final project is the text summarization tasks, there has been algorithm that help ease the process, which is using Automatic Text Summarization. However, the algorithm is not able to produce precise and accurate work results and manmade summaries still produces many errors and eliminate the automation aspects that exist in automatic text summarization. The solution to this problem is crowdsourcing, which is outsourcing tasks in the form of open calls online. However, there are no crowdsourcing platforms focused on text creation tasks. Meanwhile, text creation tasks require a workflow that may be different than other tasks in general. This final project examines workflows, task decomposition methods, and quality management appropriate for text creation tasks, using the case study text summarization tasks. <br />
<br />
<br />
<br />
<br />
In this final project, a crowdsourcing marketplace is created using the prototyping method. It must meet all the requirements of crowdsourcing text creation tasks. We tested two kinds of workflow and task decomposition methods, namely Partition-Map-Reduce and Worker Based Task Decomposition, as well as two kinds of quality management, namely Find-Fix-Verify and Evaluation by Experts with Real-time Support. <br />
<br />
<br />
<br />
<br />
This final project implementation resulted a crowdsourcing marketplace for text creation tasks called Ensemble. The results of functional testing show that out of fourteen functional requirements, nine have been fulfilled, while 80% -90% of workers agreed that the remaining five functional requirements have been met. The test results show that the Partition-MapReduce the workflow and task decomposition method, as well as quality management in the form of Evaluation by Expert and real-time support can be used for text creation tasks, and is the best for the case study and sample population in this final project.<p align=”justify”> |
format |
Final Project |
author |
Anglila Hapsari - NIM : 13514069 , Sekar |
spellingShingle |
Anglila Hapsari - NIM : 13514069 , Sekar Crowdsourcing Marketplace for Text Creation Tasks Case Study: Text Summarization Tasks |
author_facet |
Anglila Hapsari - NIM : 13514069 , Sekar |
author_sort |
Anglila Hapsari - NIM : 13514069 , Sekar |
title |
Crowdsourcing Marketplace for Text Creation Tasks Case Study: Text Summarization Tasks |
title_short |
Crowdsourcing Marketplace for Text Creation Tasks Case Study: Text Summarization Tasks |
title_full |
Crowdsourcing Marketplace for Text Creation Tasks Case Study: Text Summarization Tasks |
title_fullStr |
Crowdsourcing Marketplace for Text Creation Tasks Case Study: Text Summarization Tasks |
title_full_unstemmed |
Crowdsourcing Marketplace for Text Creation Tasks Case Study: Text Summarization Tasks |
title_sort |
crowdsourcing marketplace for text creation tasks case study: text summarization tasks |
url |
https://digilib.itb.ac.id/gdl/view/30856 |
_version_ |
1821995884528795648 |