Tools
Tools
Valluri, Sai SriMaha Vishnu (2021) Deep Learning and Natural Language Processing Approach for Real Estate Property Description Generation. Masters thesis, Dublin, National College of Ireland.
Preview |
PDF (Master of Science)
Download (3MB) | Preview |
Preview |
PDF (Configuration manual)
Download (935kB) | Preview |
Abstract
Real estate as an area of business has evolved over time. The development of online real estate marketing has contributed significantly to this improvement. The improvement of online real estate platforms along with increasing interest of individuals to handle their own sales has presented new opportunities for introducing innovative solutions into this field. In this study, exploration of the use of deep learning architectures such as convolutional neural networks and LSTM networks has been done in an attempt to create an image captioning and image tagging architecture for online real estate platforms. As a part of this, experiments have been done to understand and compare the effectiveness of attention based image captioning architectures as compared to traditional image captioning methods. The experiments have been evaluated on the basis of BLEU scores, accuracy and loss metrics. The image captioning model has showed a BLEU-1 score of 0.5456 while the accuracy of the image classification model was noted to be 76.86%.
| Item Type: | Thesis (Masters) |
|---|---|
| Subjects: | Q Science > QA Mathematics > Electronic computers. Computer science T Technology > T Technology (General) > Information Technology > Electronic computers. Computer science Q Science > QA Mathematics > Computer software T Technology > T Technology (General) > Information Technology > Computer software H Social Sciences > HD Industries. Land use. Labor > Specific Industries > Property Industry |
| Divisions: | School of Computing > Master of Science in Data Analytics |
| Depositing User: | Clara Chan |
| Date Deposited: | 15 Dec 2021 11:40 |
| Last Modified: | 15 Dec 2021 11:40 |
| URI: | https://norma.ncirl.ie/id/eprint/5232 |
Actions (login required)
 |
View Item |