a class of computational experiments that use random sampling and statistical simulation techniques to analyse the behaviour of mathematical or statistical models.
Econometrıcs I (ENG) — Ünite 8 Soru-Cevap
Econometrıcs I (ENG) (IKT325U) soru-cevapları.
What are Monte Carlo studies?
What is the equation defining the AR(p) process?
Who proposed the theory based model selection approach?
D. F. Hendry
Why is it advisable to use the unmodified Schwartz criterion for model selection when the primary goal is to maximize the probability of correctly selecting the true model?
This is because we cannot definitively determine if the coefficient of the highest lag is precisely zero or merely close to zero. In the latter scenario, the modified Schwartz criterion can be significantly inferior to the unmodified form.
For Hendry’s methodology and the purpose of model selection, what kind of a model is preferable to choose (except for forecasting purposes)?
A model that is slightly larger (thus encompassing the true model) rather than one that is too small and misspecified.
For Hendry’s methodology and the purpose of model selection, why is it not preferable to choose a model that is slightly larger for forecasting purposes?
Extraneous regressors in the model reduce the precision of estimates and result in poorer forecasts, whereas the bias introduced by excluding regressors with small coefficients tends to be comparatively small.
According to figures 8.11 and 8.12, in the case of the criteria PRESS and AIC, what is the estimation of the true lag order influenced by?
kurtosis
According to figure 8.11 and 8.12, when dealing with small sample sizes and error distributions that are not heavy-tailed (corresponding to higher degrees of freedom, k), what is advisable to utilize for selecting the lag order?
AIC
According to figure 8.11 and 8.12, when dealing with heavy-tailed error distributions, which tests demonstrate strong performances?
both PRESS and sequential F tests
According to figure 8.11 and 8.12, for large sample sizes, which are the most reliable criteria to employ?
SC and HQC
What are the other names used for "big data analytics" in the terminology of analytic methods?
“large-volume” or “large-data-set analytics” (7%)
“advanced analytics” (12%), or “analytics” (12%)
“data warehousing” (4%)
“data mining” (2%),
“predictive analytics” (2%)
other unique names (43%)
According to Witten and Frank (2005) what is the growth rate of the quantity of data stored in global databases?
The quantity of data stored in global databases doubles every 20 months (Witten and Frank, 2005)
What is "data mining"?
The exploration of data in order to uncover patterns.
What are the three dimensions of data mining?
Volume, variety, and velocity
What are the reasons for the need to analyse big data?
Competitive edge, new opportunities, improved analytical capabilities
What is the difference between statistis/econometrics and machine learning?
In statistics and econometrics, we begin with a theory and then gather data to assess its validity.
On the other hand, machine learning starts with the data and searches for patterns or theories that fit the data.
What was the focus of Senoussi’s study (2021)?
Selection of variables in the context of a nonlinear relationship between inflation and economic growth.
Where did Senoussi obtain the data for his study in 2021?
World Bank database
Which variables did Senoussi use in his study in 2021?
Log of Average GDP per Capita (Constant 2010 US Dollar)
Average Life Expectancy at Birth
Average Primary School Enrollment
Domestic Credit Provided by Financial Sector (% of GDP)
Standard Deviation of the Inflation Rate
Total Area of the Country
Average Rate of Population Growth
Average Rate of Exports of Goods and Services (% of GDP)
Average Rate of Imports of Goods and Services (% of GDP)
Average Urban Population Growth
Labour Force
Agriculture Value Added (% of GDP)
Industry Value Added (% of GDP)
Services Value Added (% of GDP)
What is the most popular programming language used for data mining/data analysis?
R