ASPECT BASED SUMMARIZATION FOR INDONESIAN ONLINE MARKETPLACE REVIEWS
Most online marketplaces in Indonesia provide review or feedback feature in order to enhance customer's satisfaction. The feature can be used by customers to give <br /> <br /> <br /> <br /> their opinion about certain sellers as well as reference for other custo...
Saved in:
| Main Author: | |
|---|---|
| Format: | Final Project |
| Language: | Indonesia |
| Online Access: | https://digilib.itb.ac.id/gdl/view/24873 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Institut Teknologi Bandung |
| Language: | Indonesia |
| Summary: | Most online marketplaces in Indonesia provide review or feedback feature in order to enhance customer's satisfaction. The feature can be used by customers to give <br />
<br />
<br />
<br />
their opinion about certain sellers as well as reference for other customers before doing transaction. However, there is a large number of unstructured opinions and <br />
<br />
<br />
<br />
every opinion can discuss one or more aspects. Therefore, the main purpose of this final project is to discover the best method in developing aspect based summarization system for reviews in Bahasa Indonesia. <br />
<br />
<br />
<br />
Aspect based summarization system in this final project consists of 4 steps, those are data crawling and filtering, sentence preprocessing, aspect and sentiment <br />
<br />
<br />
<br />
classification, and opinion summarizing. Aspect and opinion classification can be viewed as multilabel classification problem. The problem can be solved using <br />
<br />
<br />
<br />
Binary Relevance technique by developing 7 different classifiers. Each classifier is implemented using rule based and machine learning algorithms, such as Naive <br />
<br />
<br />
<br />
Bayes Classifier and Support Vector Machine. There are 7 aspects that will be used in this final project, namely goods quality, goods accuracy, service, communication, delivery, packaging, and price. <br />
<br />
<br />
<br />
The evaluation uses 2960 reviews data collected from online marketplace Tokopedia. The best method for quality, accuracy, service, communication, and delivery aspect is machine learning SVM with the average f-measures respectively are 0.7895, 0.8601. 0.8065, 0.9209, and 0.8175. Best f-measure for packaging aspect is 0.9170 and for price aspect is 0.8927. Both aspects get the best accuracy while using rule based approach in classifying aspect and sentiment. |
|---|