移动彩信相册提取,炼宝专家txt全集下载,商丘肯德基
I. Introduction: states the objectives of the project 介绍：项目的目标
在2016美国总统竞选，每个状态取决于两个主要因素：网上投票和各党派的选民的性别。
这个问题对美国的人们来说很重要，世界上其他的人们以及2016个总统候选人本身。2016个美国总统竞选的结果与美国的经济、健康、政治和安全有关。鉴于美国的全球地位，导致世界经济的领先地位，2016年代的选举结果也与世界的经济和政治有关。对于候选人本身，重要的是要了解选举结果和不同的因素之间的关系，以预测在下一个州的选票，在已完成投票的州的数据。 2016年美国总统竞选活动仍是一个持续的过程，直到十一月，2016年是令人兴奋的每天晚上投票星期二之前建立一个数据驱动的多元回归模型。到目前为止，2016年美国总统竞选没有多元回归模型建造。车型将在今年改变的一年，因为票数决定因素将发生变化。在一些年来，恐怖主义可能是一个问题;在其它年份，经济可能是一个问题。因此，这将是不科学使用旧的2012美国总统竞选模型来预测2016年竞选的结果。
The objective of this paper is to investigate the question on how the final statewide votes for Republican and Democratic candidates, in the 2016 US president campaign, of each state depend on two major explanatory factors: the online polls and the gender of voters for each party.
This question is important for the people in the US, the people in the rest of the world, and the 2016 US president candidates themselves. The outcome of the 2016 US president campaign is related to the economy, health, politics and safety to the US. Given the global status of the US, leading the economy of the world, the outcome of 2016’s campaign is also related to the economy and the politics of the world. For the candidates themselves, it is important to understand the relation between the election results and different factors, in order to forecast the votes in the next state from the data in the states that have finished voting. The 2016 US president campaign is still an ongoing process till November, 2016. It is exciting to build a data-driven multiple regression model before every voting night on Tuesdays. No multiple regression models were built so far for the 2016 US president campaign. Models will change year by year because the deciding factors of the votes will change. In some years, the terrorism might be an issue; in other years, the economy might be an issue. Therefore, it will be unscientific to use the old 2012 US president campaign model to predict the outcomes of 2016 campaign.
This paper is going to answer this question by building a multiple regression model of the final statewide polls for the states, that have finished voting, to forecast the final votes of Republican and Democratic winners of the New York states, which will happen on 19/04/2016, through statistical means in Econometric business forecasting. This paper is organised as the follows. In Section II, a literature review of past statistics modelling of US presidential campaign is presented. In Section III, the data of the multiple regression model is collected and its descriptive statistics is presented in tables and figures. In Section IV, the multiple regression model of this paper is explained. In Section V, the parameters in the model is estimated and the hypothesis is tested. In Section VI, a brief summary is provided along with the business forecast of the 2016 US presidential campaign results in the New York state.
II. Literature review文献综述
Several factors can influence the outcome of the 2016 US president campaign. Among them, the major factors are the party of the candidate (Republican or Democratic), how the candidate is going to create new jobs, how the candidate is going to boost the US economy, how the candidate is going to improve the education, how the candidate is to solve the terrorism problem and scandals (L.J. Sabato et al, 2015). From the CNN exit polls, the voters from the Republican side and the Democratic side weigh differently on these factors state by state (CNN politics, 2016).
Past work has been done on building multiple regression models for the outcomes of votes in US president campaigns. A multiple regression model is built to relate the US presidential elections to the US economy (R.K. Roth, 2012). In 2012, a multiple regression model discovered that the unemployment rate as a major factor in the 2012 US presidential campaign (P. Sinha et al, 2012). In this paper, a different approach is adopted to predict the statewide votes from the online polls. Among all the online polls, the Real Clear Politics (RCP) averaged polls considered all the online up to date polls (Real Clear Politics, 2016). This is similar to Nate Silver’s regression model based on polls at the state level for the 2012 campaign (N. Silver, 2013). Nate Silver used the Public Policy Polling (PPP) (Public Policy Polling, 2016), however the RCP averaged polls will be used in my model because RCP averaged polls have included nearly all the online polls, including the PPP results. From the statewide exit polls on CNN exit polls, the votes in each state are sensitive to the percentage of male voters during the exit polls (CNN politics, 2016).
III. Data数据
The 2016 US president campaign is still an ongoing process till the end of 2016. The source of the data being used to build a multiple regression model is explained as the follows. For the states that have finished the voting process, either through caucuses or primaries (K. Lewis, 2016), the votes for the winner, for both the Republican and the Democratic sides, are collected state by state in percentages from the CNN politics (CNN politics, 2016). The statewide votes form the data of the dependent variable. The RCP averaged votes for the winner, either from the Republican or from the Democratic side, in each voted states are collected in percentages from the RCP (Real Clear Politics, 2016). The statewide male percentage, both from the Republican and the Democratic sides, during the exit polls are collected state by state from the CNN exit polls (CNN politics, 2016). Both the RCP averaged votes for the winner of the states, denoted as RCP_R and RCP_D for the Republic and the Democratic, and the statewide male percentage, denoted as MEN_R and MEN_D for the Republic and the Democratic, during the exit polls are regarded as explanatory variables in this model. Besides that, a dummy variable (I) is introduced to label the Republican party (I = 1) and the Democratic party (I = 0). 34 observations are made from the winners of 34 voted states either on the Republican and Democratic sides. Please see the attached Excel file for the data.
A scattered plot of votes (denoted as VOTE) versus RCP_R and MEN_R, or RCP_D and MEN_D is shown in Figure 1. From Figure 1, we see that the data points almost lied on a plane, indicating that a multiple linear regression model might be a good fit for the data.
Figure 1: 3D scatter plot of statewide votes of winners VOTE (in %) versus statewide RCP average (in %) and percentage of men in the statewide exit polls. 34 observations in total.
The descriptive statistics of the collected data are presented in Table 1.
mean mediam st. dev min max
VOTE 52.0235 50 14.3252 27.6000 86.1000
RCP_D 55.2579 56.3000 10.2727 37.3000 84.5000
MEN_D 42.5263 43 2.5684 36 46
RCP_R 35.7533 37.2000 6.6662 20.3000 45.3000
MEN_R 50.9333 51 1.4864 49 53
 |
