在經典的社群網路分析問題:節點分類與連結預測中，如何有效提升準確率是分析者所努力的目標，
學習出更有用的嵌入學習表示是可以被改進的地方，本論文提出透過加入全域(Global)的資訊，改善了一般社群網路嵌入學習(Network Embedding)無法抓到全域資訊的缺點，本研究提出了三種策略：基於社群壓縮策略、基於社交圈壓縮策略、基於星狀壓縮策略，透過不同的角度提供全域的資訊，使預測準確率能上升。
本研究也探討了基於不同的網路嵌入學習方法及不同的資料集下預測結果的表現差異，結果顯示，本論文所提出無論於何種壓縮策略，只要經由壓縮策略壓縮後網路圖所學習出的嵌入向量，串接回原始網路圖後形成特徵矩陣，使用特徵矩陣訓練模型並預測結果，本論文提出的方法預測表現都能比只使用社群網路嵌入學習方法好。
在三種壓縮策略中又以基於社群壓縮策略提供最廣義的全域資訊，表現結果最佳，基於社交圈壓縮策略則是介於廣義與狹義間的方法，表現結果次之，而基於星狀壓縮策略表現結果最不優異，認為基於星狀壓縮策略提供了較狹義的全域資訊，但三種策略都較原始未加入策略下的方法好。

The main academic contribution of this thesis effectively introduce global information into network embedding learning and acquire a significant improvement in the experiment of node classification and link prediction.
Specifically, the original network embedding learning which generates a node sequence through random walk only considers its local information (immediate neighbors).
In addition, global information such as node communities and social circles is being used as compressing nodes and forms multi-level super graphs, so that network embedding learning can be also learned from the global information.
The Compression-based Feature Representation Learning is a general-purpose architecture for global information in various of applications network embedding learning models such as deepwalk, node2vec, LINE, and struc2vec. The experimental result shows the accuracy of deepwalk and node2vec can be significantly improved from 0.3 to 0.6.%by about 15% - 20%.

[1] Blondel, V. D., Guillaume, J.-L., Lambiotte, R., and Lefebvre, E. (2008). Fast unfolding of communities in large networks. Journal of statistical mechanics: theory and experiment,2008(10):P10008.
[2] Dong, Y., Chawla, N. V., and Swami, A. (2017). metapath2vec: Scalable representation learning for heterogeneous networks. In Proceedings of the 23rd ACM SIGKDD international conference on knowledge discovery and data mining, pages 135–144. ACM.
[3] Gao, H. and Huang, H. (2018). Deep attributed network embedding. In IJCAI, pages 3364–3370.
[4] Grover, A. and Leskovec, J. (2016). node2vec: Scalable feature learning for networks. pages 855–864.
[5] Hartigan, J. A. and Wong, M. A. (1979). Algorithm as 136: A k-means clustering algorithm. Journal of the Royal Statistical Society. Series C (Applied Statistics), 28(1):100–108.
[6] Ma, J., Cui, P., Wang, X., and Zhu, W. (2018). Hierarchical taxonomy aware network embedding.
In Proceedings of the 24th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining, pages 1920–1929. ACM.
[7] Mikolov, T., Chen, K., Corrado, G., and Dean, J. (2013). Efficient estimation of word representations in vector space. arXiv preprint arXiv:1301.3781.
[8] Perozzi, B., Al-Rfou, R., and Skiena, S. (2014). Deepwalk: Online learning of social representations. pages 701–710.
[9] Ribeiro, L. F., Saverese, P. H., and Figueiredo, D. R. (2017). struc2vec: Learning node representations from structural identity. In Proceedings of the 23rd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, pages 385–394. ACM.
[10] Tang, J., Jiang, B., Zheng, A., and Luo, B. (2012). Graph matching based on spectral embedding with missing value. Pattern Recognition, 45(10):3768–3779.
[11] Tang, J., Qu, M., Wang, M., Zhang, M., Yan, J., and Mei, Q. (2015). Line: Large-scale information network embedding. In Proceedings of the 24th international conference on world wide web, pages 1067–1077. International World Wide Web Conferences Steering
Committee.
[12] Watts, D. J. and Strogatz, S. H. (1998). Collective dynamics of‘small-world’networks. nature, 393(6684):440.
[13] Ye, Z., Zhao, H., Zhang, K., and Zhu, Y. (2019). Multi-view network representation learning algorithm research. Algorithms, 12(3):62.