时间序列代写 | Regression Analysis Homework 3

本次美国代写主要为统计学时间序列相关的assignment

指示：

下载名为hw3_data.csv的数据

30分–回归分析
5分执行EDA（数字，视觉，描述性）
5分什么是奥肯定律？我们的数据集是否符合我们的期望？为什么或者为什么不？
5分创建一个以百分比显示GDP变化的列，并以一个百分比变化显示UE费率的列。对于日期4/1/1948 GDP更改= 2.568，UE差异= 0.5
5分分为训练/测试（）并创建散点图，以验证奥肯定律中的负相关关系。
在训练集上拟合线性模型。使用您的数据说明奥肯定律解释系数。
5分我们如何使用此模型来预测未来的GDP美元水平？我们需要什么，带来的局限性是什么？

不幸的是，我们还没有学会如何为多元时间序列拟合Box-Jenkins模型。因此，我们将遵循在第3周幻灯片中找到的模型识别程序，以对数据集中的经济时间序列进行EACH。从原始时间序列数据开始：GDP（美元）和UE率

30分– Box Jenkins GDP模型
7分执行差异或变换，直到您可以从视觉和统计上证明训练集的平稳性为止。总结对时间序列的更改。
7分创建和解释ACF和PACF图
7分在训练集上拟合适当的Box Jenkins模型
7分检查残差（散点图，直方图，ACF，portmanteau测试）。总结残差。
7分在测试集上打印模型评估指标MAE，MSE，MAPE，sMAPE
35分– Box Jenkins失业率模型
7分执行差异或变换，直到您可以从视觉和统计上证明训练集的平稳性为止。总结对时间序列的更改。
7分创建和解释ACF和PACF图
7分在训练集上拟合适当的Box Jenkins模型
7分检查残差（散点图，直方图，ACF，portmanteau测试）。给出残差的摘要。
7分在测试集上打印模型评估指标MAE，MSE，MAPE，sMAPE
10分-评估使用误差指标作为证据比较两个模型的性能

Instructions:

Download the data named hw3_data.csv

1. 30 pts – Regression Analysis
   1. 5 pts Perform EDA (numerical, visual, descriptive)
   2. 5 pts What is Okun’s Law? Does our dataset fit what we would expect to see? Why or why not?
   3. 5 pts Create a column that shows GDP change as a percentage and a column that shows UE rate as a change in percentage points. For date 4/1/1948 GDP change = 2.568, UE diff = 0.5
   4. 5 pts Split into Train/test ( ) and create a scatterplot, verifying the negative relationship in Okun’s law.
   5. Fit a linear model on the Training set. State Okun’s law using your data e.g. interpret the coefficients.
   6. 5pts How can we use this model to predict level of GDP dollars in the future? What will we need, what are the resulting limitations?

Unfortunately, we have not yet learned how to fit Box-Jenkins models for multivariate time series. Thus, we will follow the Model identification procedure found in Week 3 slides for EACH of the economic time series in the dataset. Start with the original time series data: GDP in dollars and UE rate

2. 30 pts – Box Jenkins model for GDP
   1. 7 pts Perform differencing or transformations until you can visually and statistically prove stationarity on training set. Summarize the changes to your time series.
   2. 7 pts Create and interpret ACF and PACF plots
   3. 7 pts Fit the appropriate Box Jenkins model on training set
   4. 7 pts Check residuals (scatter plot, histogram, ACF, portmanteau tests) Give a summary of your residuals.
   5. 7 pts Print model evaluation metrics MAE, MSE, MAPE, sMAPE on test set
3. 35 pts – Box Jenkins model for Unemployment rate
   1. 7 pts Perform differencing or transformations until you can visually and statistically prove stationarity on training set. Summarize the changes to your time series.
   2. 7 pts Create and interpret ACF and PACF plots
   3. 7 pts Fit the appropriate Box Jenkins model on training set
   4. 7 pts Check residuals (scatter plot, histogram, ACF, portmanteau tests)  Give a summary of your residuals.
   5. 7 pts Print model evaluation metrics MAE, MSE, MAPE, sMAPE on test set
4. 10 pts – Evaluation Compare the two models’ performance using the error metrics as evidence