在2018年的選舉中，六都的國民黨候選人初期民調都不高，藉由高雄市長候選人的名氣提升，除了在高雄勝選外，也使國民黨獲得六都中半數的執政權。根據因素分析，可知六都民調中有兩個主要因素，較大的因素為韓流，這股潮流以逐漸上揚的方式貫穿了選戰後期。
我們以Galam模型擬合六都民調數據，藉由新聞及社會事件決定擬合時段，進而預測投票結果。同時我們也根據韓流現象提出另一新模型，其中增加廣域磁場變數可以影響所有選民，稱為磁場模型，此模型改善了Galam模型中僅在偶數組織中，相同票數的情況下才能調整參數的問題。
以Galam模型及磁場模型預測六都選舉結果，我們發現使用小組織所得到的預測誤差較小。若有影響民調的新聞或事件發生，在選取時段上，越接近投票日的時段預測結果越好。

In 2018 Taiwanese municipal elections, the six Kuomintang special municipality candidates didn’t have high voter support at the early stage. The Kuomintang candidate at Kaohsiung gradually became famous and won voter support. This trend helped him to win the Kaohsiung election and also helped Kuomintang to win three elections in Taiwan’s six special municipalities. By applying the factor analysis to the polling data, we find there were two major factors for special municipality polls, and the larger one can be identified as the “Han current”, which is almost a monotonically increasing time series in the later stage of the elections.
We use the Galam model to fit the polling data and predict the election results. The time period for the data fitting is selected according to relevant news and social events. Inspired by the “Han current” collective phenomena, we also propose a new model, the magnetic field model, in which a global magnetic field influences all the voters. The tunable parameter in the Galam model only acts on the term with the same number of voter support in an even-voter party, while the magnetic field parameter acts on every term in the model.
When the Galam model and the magnetic field model is used to predict election results, we find that adopting small parties of voters will increase the accuracy of prediction. If there are several news scandals or social events affecting the polls, it is better to choose the time period starting from the news or social events closer to the polling day.
 

第一章 緒論 1
第二章 研究方法 3
2.1 展透 (Percolation) 3
2.1.1 一維展透模型 3
2.1.2 展透重整化群方法 (Percolation Renormalization) 5
2.1.3 幽靈晶格(Ghost Site) 6
2.2 Galam Model 6
2.3.1 Galam Model在兩黨選舉上的運作 7
2.3.2 相對低支持度的勝利者 8
2.3.3三黨政治下的Galam模型 9
2.3 磁場模型 10
2.3.1 兩黨的磁場模型 10
第三章 相關矩陣 (Correlation matrix) 11
3.1 因素分析 (Factor Analysis) 12
3.2 2018國民黨六都市長選舉民調支持度數據因素分析 14
第四章 中華民國市長選舉 16
4.1 高雄選舉 16
4.1.1高雄市長選舉以Galam模型模擬 20
4.1.2 韓國瑜潮流的效益與磁場效應 25
4.2 新北選舉 29
4.3 桃園選舉 36
4.4 台中選舉 42
4.5 台南選舉 47
4.6 台北選舉 53
第五章 結論 67
參考文獻Reference 68

[1] C. Castellano, S. Fortunato, V. Loreto, Statistical physics of social dynamics, Rev. Mod. Phys. 81, 591 (2009).
[2] S. Galam, Sociophysics: A Physicist's Modeling of Psycho-political Phenomena (Springer, 2012).
[3] P. Sen, B. K. Chakrabarti, Social Physics: An Introduction (Oxford Univ. Press 2014).
[4] S. Galam, Majority rule, hierarchical structures, and democratic totalitarianism: A statistical approach, J. Math. Psychol. 30, 426 (1986).
[5] D. Stauffer, A. Aharony, Introduction to Percolation Theory (Taylor & Francis, 1994).
[6] S. R. Broadbent, J. M. Hammersley, Percolation processes: I. Crystals and mazes, Math. Proc. Cambridge Phil. Soc., 53, 629 (1957).
[7] S. Gekle, L. Peliti, S. Galam, Opinion dynamics in a three-choice system, The Eur. Phys. J. B 45, 569 (2005).
[8] J. Fernabdez-Gracia, K. Suchecki, J. J. Ramasco, M. S. Miguel, V. M. Eguiluz, Is the Voter Model a Model for Voters, Phys. Rev. Lett. 112, 158701 (2014).
[9] S. Galam, Social paradoxes of majority rule voting and renormalization group, J. Stat. Phys. 61, 943 (1990).
[10] S. Galam, S Moscovici, Towards a theory of collective phenomena: Consensus and attitude changes in groups, Eur. J. Soc. Psychol. 21, 49 (1991).
[11] S. Galam, Real space renormalization group and totalitarian paradox of majority rule voting, Physica A 285, 66 (2000).
[12] S. Galam, Minority opinion spreading in random geometry, Eur. Phys. J. B 25, 403 (2002).
[13] S. Galam, Contrarian deterministic effects on opinion dynamics:”the hung elections scenario”, Physica A 333, 453 (2004).
[14] S. Galam, Local dynamics vs. social mechanisms: A unifying frame, Europhys. Lett. 70, 705 (2005).
[15] S. Galam, F. Jacobs, The role of inflexible minorities in the breaking of democratic opinion dynamics, Physica A 381, 366 (2007).
[16] S. Galam, Sociophysics: A review of Galam models, Int. J. Mod. Phys. C 19, 409 (2008).
[17] S. Galam, The Trump phenomenon: An explanation from sociophysics, Int. J. Mod. Phys. C 31, 1742015 (2017).
[18] T. Cheon, S. Galam, Dynamical Galam model, Phys. Letts. A 382, 1509 (2018).