Tools
Tools
Kumari, Apurva (2022) Exploration of the Most Preferred Social Media for the Fashion Business Practices. Masters thesis, Dublin, National College of Ireland.
Preview |
PDF (Master of Science)
Download (1MB) | Preview |
Preview |
PDF (Configuration manual)
Download (1MB) | Preview |
Abstract
This study looks into the problem to find out how social media affects the fashion industry. In the last ten years, social media have come a long way. Companies are using social media as a way to market their products. Social media is the best and cheapest way for the fashion industry, which is one of the industries with the most growth, to talk to people. Text processing is one of the most natural applications for machine learning and deep learning models. There are, however, few research that combine the two. This study uses a combination of sentiment analysis and text mining to find out how customers behave and how happy they are with different fashion businesses on social media, where customers are very active in sharing their opinions in real time. This study uses Twitter data in many ways, including retrieval, cleaning, feature selection, and classification using three different machine learning algorithms (Naive Bayes, XGBoost, and SVM) and one Deep ML technique (BERT). After the raw data has been cleaned up and the key features for a classification algorithm have been extracted, classification and model validation are done. This study shows that, compared to other classification models, the BERT model has the highest classification accuracy and the highest weighted value of precision, with 99.0% accuracy. The results of this study say that almost 61% of the texts collected had positive things to say about different fashion labels. When it comes to the number of tweets per country, Kampala, Uganda is at the top of the list.
Actions (login required)
 |
View Item |