Citation network analysis with deep learning
Heterogeneous information networks (HINs) can be found everywhere in real-world applications. At the same time, node embedding has been regarded as a useful tool to mine and learn from networked data. As a result, it is of interest to develop HIN embedding methods. However, the heterogeneity in HINs...
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Final Year Project (FYP) |
| Language: | English |
| Published: |
2019
|
| Subjects: | |
| Online Access: | http://hdl.handle.net/10356/77382 |
| _version_ | 1826109513809264640 |
|---|---|
| author | Zhao, Xinghe |
| author2 | Chen Lihui |
| author_facet | Chen Lihui Zhao, Xinghe |
| author_sort | Zhao, Xinghe |
| collection | NTU |
| description | Heterogeneous information networks (HINs) can be found everywhere in real-world applications. At the same time, node embedding has been regarded as a useful tool to mine and learn from networked data. As a result, it is of interest to develop HIN embedding methods. However, the heterogeneity in HINs introduces not only rich information but also potentially incompatible semantics, which poses special challenges to embedding learning in HINs. With the intention to preserve the rich yet potentially incompatible information in HIN embedding, we propose to study the problem of comprehensive transcription of heterogeneous information networks. The comprehensive transcription of HINs also provides an easy-to-use approach to unleash the power of HINs, since it requires no additional supervision, expertise, or feature engineering. To cope with the challenges in the comprehensive transcription of HINs, we studied the HEER algorithm, which embeds HINs via edge representations that are further coupled with properly-learned heterogeneous metrics. To verify the efficacy of HEER, we conducted experiments on real-world datasets DBIS with an edge reconstruction task and also run simulations on several case studies to fully understand the strengths and weakness of HEER. Besides, we explored a newly proposed model, which is to enhance HEER by adding Meta-path information to improve the performance. Experiment results demonstrate the effectiveness of the new proposed model and the utility of edge representations and heterogeneous metrics. |
| first_indexed | 2024-10-01T02:19:27Z |
| format | Final Year Project (FYP) |
| id | ntu-10356/77382 |
| institution | Nanyang Technological University |
| language | English |
| last_indexed | 2024-10-01T02:19:27Z |
| publishDate | 2019 |
| record_format | dspace |
| spelling | ntu-10356/773822023-07-07T16:42:49Z Citation network analysis with deep learning Zhao, Xinghe Chen Lihui School of Electrical and Electronic Engineering DRNTU::Engineering::Electrical and electronic engineering Heterogeneous information networks (HINs) can be found everywhere in real-world applications. At the same time, node embedding has been regarded as a useful tool to mine and learn from networked data. As a result, it is of interest to develop HIN embedding methods. However, the heterogeneity in HINs introduces not only rich information but also potentially incompatible semantics, which poses special challenges to embedding learning in HINs. With the intention to preserve the rich yet potentially incompatible information in HIN embedding, we propose to study the problem of comprehensive transcription of heterogeneous information networks. The comprehensive transcription of HINs also provides an easy-to-use approach to unleash the power of HINs, since it requires no additional supervision, expertise, or feature engineering. To cope with the challenges in the comprehensive transcription of HINs, we studied the HEER algorithm, which embeds HINs via edge representations that are further coupled with properly-learned heterogeneous metrics. To verify the efficacy of HEER, we conducted experiments on real-world datasets DBIS with an edge reconstruction task and also run simulations on several case studies to fully understand the strengths and weakness of HEER. Besides, we explored a newly proposed model, which is to enhance HEER by adding Meta-path information to improve the performance. Experiment results demonstrate the effectiveness of the new proposed model and the utility of edge representations and heterogeneous metrics. Bachelor of Engineering (Electrical and Electronic Engineering) 2019-05-28T02:32:46Z 2019-05-28T02:32:46Z 2019 Final Year Project (FYP) http://hdl.handle.net/10356/77382 en Nanyang Technological University 50 p. application/pdf |
| spellingShingle | DRNTU::Engineering::Electrical and electronic engineering Zhao, Xinghe Citation network analysis with deep learning |
| title | Citation network analysis with deep learning |
| title_full | Citation network analysis with deep learning |
| title_fullStr | Citation network analysis with deep learning |
| title_full_unstemmed | Citation network analysis with deep learning |
| title_short | Citation network analysis with deep learning |
| title_sort | citation network analysis with deep learning |
| topic | DRNTU::Engineering::Electrical and electronic engineering |
| url | http://hdl.handle.net/10356/77382 |
| work_keys_str_mv | AT zhaoxinghe citationnetworkanalysiswithdeeplearning |