Social media has over the years partly become a platform to express opinions and discuss current events. Within the field of Computer Science, Twitter has been used both as the basis for political analysis - for example using sentiment analysis to predict election results - and within the field of cluster analysis, where the question of how to best design and use an algorithm to extract topics from tweets has been studied. The ClusTop algorithm is specifically designed to cluster tweets based on topics. This paper aims to explore whether it is possible to (a) use an implementation of the ClusTop algorithm to identify topics connected to tweets about Trump and Clinton just before the American 2016 election, and (b) distinguish between the topics used in connection with a specific candidate in states where they won versus states where they lost the election. The problem is approached through the method of a controlled experiment where the data collected from Twitter is divided into groups and run through the ClusTop algorithm. The topics are then compared to draw tentative conclusions about their validity as a basis for election prediction. The study finds that it is indeed possible to adapt the ClusTop algorithm to use with tweets and geolocation to identify different topics, thus confirming the usefulness of the algorithm. In addition to this, the study confirms that manually examining the words used within the topics makes it possible to see differences between them. The work thereby places itself in the tradition of exploring how Twitter can be used for election prediction by being one of the first studies to look at clustering as a way of approaching the problem.

With the rise of social media, people have gained a platform to express opinions and discuss current subjects with others. This thesis investigates whether a simple sentiment analysis — determining how positive a tweet about a given party is — can be used to predict the results of the Swedish general election and compares the results to betting odds and opinion polls. The results show that while the idea is an interesting one, and sometimes the data can point in the right direction, it is by far a reliable source to predict election outcomes.

中華民國第十任總統選舉結果於民國八十九年三月十八日揭曉，這場眾所矚目的選舉終告落幕，然而對選舉研究工作者而言卻是新的開始。選舉預測居選戰中重要的一環，也是研究選舉的學者關心的問題，更提供了一個驗證選民投票行為理論的絕佳機會，近來國內相關論述已有相當成果。但由於它在投票結束，便有答案，其挑戰程度不言而喻。因此，如何結合理論、方法及事實三者為一體的努力，對選舉預測更是別具意義。 本篇研究之範圍，是以公元2000年總統大選為例，對選舉預測工作做更深層的探討，且檢驗邏輯斯預測模型（Logistic Regression Model）及模糊統計（Fuzzy Statistics）分析在本次總統選舉的預測力，考量本次總統選舉中各項可能影響選情的因素，進一步建構選舉預測模式，然而兩種預測模式的初步預測結果並不佳，經過棄保效應的可能性調整後，預測誤差已大幅降低，其中模糊統計（Fuzzy Statistics）分析預測結果經棄保效應調整後，與實際開票結果相當接近，因此與邏輯斯預測模型相較，模糊統計分析的應用對未表態選民投票意向的預測力較佳。一套完整的選舉預測模型研究，應包含問卷設計、抽樣訪問、資料處理、加權除錯、模型設計與預測評估等整套研究流程，然而在本次總統大選中，由於三強激戰，影響選情因素相當複雜，最後此兩種選舉預測模式皆無法獲致精確的預測結果。因此，我們期待選舉預測模型的建構，能突破主客觀環境的侷限，進一步達到「準」與「穩」的要求。 / With the successful staging of the 2000 presidential elections in Taiwan, scholars have been presented with a new opportunity to test their theories. Electoral predictions are an important field within the study of elections and have been among the most keenly studied questions over the past few years. Unlike many other research topics, there is an absolute standard for election predictions: the election results. Thus, combining theory, methodology, and facts to obtain a meaningful result is no simple task. This thesis attempts to predict the 2000 presidential election using both a logistic regression model and a fuzzy statistics model. After constructing models which includes all kinds of different variables that might influence the electoral outcome, we find that neither the logistic regression model nor the fuzzy statistics model is particularly accurate. However, after accounting for the effects of strategic voting, model error decreases dramatically. In particular, after including provisions for strategic voting, the fuzzy statistics model is improved to the point that its predictions are extremely close to the actual outcome. Thus, we show that the fuzzy statistics model is superior to the logistic regression model in analyzing the vote choices of undecided voters. Research on electoral predictions should include such aspects as questionnaire design, sampling, interviewing, data processing, weighting, data cleaning, model design, and evaluation of the prediction. However, because this election featured a particularly intense three way race, the factors affecting the electoral outcome were both numerous and intertwined in complex ways. Unfortunately, it is impossible to evaluate our electoral predictions of the two models precisely. We hope that in the future, election prediction models will be able to break through these environmental limitations and achieve more accurate and stable predictions.

The diploma thesis "Estimation of the election results by respondents and its usage" discuss the methods of predicting the election results based on election expectations of individuals and shows how people form their expectations and how exact these expectations are. First short summary of existing methods of election results predictions and its limitations is presented - it also deals with pre-election surveys and its ambitions to predict the election results. The rest of the thesis focuses only on the prediction of the election results based on election expectations of individuals: prediction markets and aggregated estimations of respondents. Concept Wisdom of Crowds, from which both approaches originate, is presented together with concrete examples of application of the predictions based on opinions of prediction markets participants and respondents of the pre-election surveys. Results of the foreign studies confirm that the prediction markets predict the election results very well and with higher accuracy than the pre-election surveys. Current studies also positively evaluate the estimation of the election results done by respondents. Respondents are usually able to predict the election results, even several weeks before the elections. Last part of the thesis is based on my own quantitative...