Calculate the coefficient of determination.

	a. 0.37
	b. 0.43
	c. 0.85
	d. 0.12

.

If s_x and s_y are the standard deviations of x and y, and r_xy is the sample correlation coefficient between x and y, which of the following is the slope of a simple linear regression y = α+ βx ?

	a. rxy
	b. rxy sy/sx
	c. rxy (sy/sx)2
	d. sy/sx

.

The following data x =[18 20 22 24] and y = [76.1 77.0 78.1 79.7] are fitted with a linear regression. What is the slop of the line?

	a. 0.60
	b. 0.1
	c. 0.9
	d. 0.7

.

The sample mean and the variance for X are 0 and 1. The sample mean and the variance of Y are 2 and 2. The sample correlation is 0.7. What is the intercept of the least squares line?

	a. 2
	b. 0.99
	c. 1
	d. 0

.

The sample variance for X and Y are 1 and 4, respectively. The sample correlation is 0.5. What is the slope of the least squares line?

	a. 1
	b. 2
	c. 0.5
	d. 0.25

.

What are the consequences of heteroscedasticity in ordinary least square estimates?

	a. Heteroscedasticity results in biased parameter estimates.
	b. Heteroscedasticity leads to inefficient estimation.
	c. Heteroscedasticity leads to inconsistency.
	d. All of the above.

.

What is ANOVA used for?

	a. To determine if the means of two samples are equal
	b. To determine if the variances of two or more samples are equal
	c. To determine if the variances of more than two samples are equal
	d. To determine if the means of two or more populations are equal

.

What is the F ratio in a one-way ANOVA?

	a. MSTR/MSE
	b. MST/MSE
	c. MSE/MSTR
	d. MSE/MST

.

When there are only two groups for a one-way ANOVA, what is the relationship between F-test and t-test?

	a. F = t
	b. F = t2
	c. F > t
	d. F < t

.

Which of the following assumption is NOT necessary for conducting a one-way ANOVA comparing two population means?

	a. The variances of the response variables of the two populations are equal.
	b. The values of the response variables are normally distributed.
	c. The samples from the two populations are randomly selected, independent samples.
	d. The sample sizes of two populations are equal.

.

Which of the following is NOT a component of a summary table for ANOVA?

	a. F-statistics
	b. Degree of freedom
	c. Correlation coefficients
	d. Sum of squares

.

Which of the following is not true about the correlation coefficient?

	a. The correlation coefficient is also known as Pearson’s r, named after its inventor Karl Pearson.
	b. The value of a correlation coefficient computed from a sample always lies between -1 and +1.
	c. A significant correlation indicates a causal relationship between two random variables.
	d. When a sample correlation is significant, the null hypothesis of no linear association can be rejected.

.

Which of the following statements about chi-square test is NOT true?

	a. A chi-square goodness of fit test is valid if each of the expected cell frequencies is less than five.
	b. A contingency table is a type of table in a matrix format that displays the frequency distribution of the variables.
	c. The smaller the value of the chi-square test statistic, the more likely the null hypothesis will be rejected.
	d. The statistic is zero if the observed frequencies are equal to the expected frequencies.

.

Which of the following statements is NOT true?

	a. The variance of the residuals has no influence on the uncertainty in estimating the regression coefficients.
	b. In simple linear regression, the slope of the regression line is proportional to the correlation between X and Y.
	c. The R2 for a regression of Y onto X and the R2 for the regression of X onto Y are equal.
	d. Least squares residuals are not correlated with the fitted values.

.

Which of the following values could not represent a correlation coefficient?

	a. 0.5
	b. 0.9
	c. -0.5
	d. 1.5

.

Which quantity measures the variability of the observed values of the response variable around their respective treatment means?

	a. Mean
	b. Error sum of squares
	c. Correlation
	d. None of the above

.

You conducted a 2 (gender) x 4 (treatment group) analysis of responses to 4 different treatment regiments for depression. You conducted a two-way ANOVA and found that the F-value for the corrected model is significant. Which conclusion can you make based on this analysis?

	a. There is a significant difference between at least two of the groups.
	b. There is a significant difference between genders.
	c. There is a significant difference between at least two of the four treatment regimens.
	d. There is a significant difference between all eight groups.

.

You have three groups of patients, and each group is comprised of ten males and ten females. The dependent variable is the reduction in cholesterols following a treatment. You predict that Group A will respond to treatment better than Group C, and that males will respond to treatment less than females. Which method do you use to confirm your prediction?

	a. two-way ANOVA
	b. one-way ANOVA
	c. t-test
	d. chi-square test

.

A multiple regression model has p independent variables. What is the degree of freedom for error if data contain N observations?

	a. N-p-1
	b. N-p+1
	c. N-p
	d. N-1

.

Calculate R² for a multiple regression model, with SSR = 100 and SSE = 200.

	a. 0.333
	b. 0.5
	c. 0.300
	d. 0.75

.

Compute the value of F_0.5 with 8 numerator and 19 denominator degrees of freedom.

	a. 2.12
	b. 2.48
	c. 7.50
	d. 1.56

.

Determine the value of F_0.05 with 6 numerator and 60 denominator degrees of freedom.

	a. 2.3
	b. 5.2
	c. 2.5
	d. 1.2

.

Fill in the blank. ________ is used to test the significance of multiple regression model.

	a. The overall F-test
	b. The t-test
	c. The partial F-test
	d. The chi-square test

.

In order to test for the significance of a regression model involving 5 independent variables and 40 observations, the denominator degree of freedom for the critical value of F is:

	a. 5
	b. 35
	c. 45
	d. 34

.

Let y_i be the observed values in a dataset and f_i be the modeled values from a regression. Let ymean be the mean of the observed data. Which of the following is the regression sum of squares (SSR)?

	a. ∑(yi-ymean)2
	b. ∑(fi-ymean)2
	c. ∑(fi-yi)2
	d. None of the above

.

R² is computed to be 0.79 for a multiple regression analysis on fifty independent variables. What is the appropriate interpretation of R² = 0.79?

	a. The model predicts outcomes 79% of the time.
	b. Twenty-one percent of the independent variables should be removed from the analysis.
	c. Seventy-nine percent of variations in the observed values of the dependent variable are explained by the independent variables.
	d. Only seventy-nine percent of the independent variables are significant.

.

The correlation coefficient between two random variables is 0.85, which indicates

	a. negative linear correlation
	b. some linear correlation
	c. no linear correlation
	d. perfect linear correlation

.

What is the objective of multiple regression?

	a. Quantify correlations between predictors
	b. Identify causal relationships between variables
	c. Explain variation in one variable based on variations of other variables
	d. All of the above

.

When do you need to use dummy variables in a multiple regression?

	a. When performing residual analysis
	b. When correcting for multicollinearity
	c. When qualitative variables are used in the model
	d. None of the above

.

Which factors do the adjusted coefficient of determination adjust for?

	a. Number of independent variables and sample size
	b. Number of dependent variables
	c. Sample size
	d. Significant level and sample size

.

Which is the range of a correlation coefficient r?

	a. r > 1
	b. -1≤r≤1
	c. r ≤1
	d. 0≤r≤1

.

Which of the following is always true when we add an independent variable to a multiple regression model?

	a. Adjusted coefficient of determination decreases.
	b. Unadjusted coefficient of determination increases.
	c. Adjusted coefficient of determination increases.
	d. Unadjusted coefficient of determination decreases.

.

Which of the following is NOT true about principal component analysis (PCA)?

	a. PCA transforms data to a new coordinate system, with each coordinate being referred to as a principal component.
	b. In PCA, each of the principal components is a nonlinear combination of the original variables.
	c. The principal components are arranged in order of decreasing variance.
	d. The most informative principal component is the first component.

.

Which of the following is the definition of R²? SSR is the regression sum of squares, SST is the total sum of squares, and SSE is the sum of squares of residuals.

	a. SSR/SST
	b. SSE/SSR
	c. SSE/SST
	d. SST/SSE

.

Which of the following statement is NOT true?

	a. The sum of the residuals for a least squares line is zero.
	b. R2 is equal to the square of the sample correlation between the observed values and the fitted values.
	c. R2 = SSR/SST.
	d. If the correlation rX,Y is 0, there is no relationship between X and Y.

.

You were asked to conduct a regression analysis to determine the relationship between a dependent variable and 8 independent variables, using 56 observations. You obtained the following information from the computer outputs: R Square = 0.80 and SSR = 4,280. Determine the value of F-test for the regression.

	a. 12.4
	b. 24.5
	c. 11.2
	d. 27.3

.

You were asked to perform a regression analysis performed on a sample with thirty-five observations to quantify the relationships between one dependent variable and four independent variables. You obtained R2 = 0.6 and sum of squares (SSR) is 4800. Determine F-statistics for the regression and whether the model is significant at α = 0.05 level of significance.

	a. 29.55 and the model is significant
	b. 11.25 and the model is significant
	c. 2.5 and the model is not significant
	d. 7.5 and the model is not significant

.

You were asked to perform a regression analysis performed on a sample with forty-five observations to quantify the relationships between one dependent variable and four independent variables. You obtained R² = 0.8 and sum of squares (SSR) is 680. Determine the degrees of freedom and F-statistics for the regression.

	a. 4 and 45
	b. 4 and 40
	c. 5 and 20
	d. 2 and 10

.

Fill in the blank. A variable Z, whose value is Z = X₁X₂, is introduced to a general linear model in order to account for potential ________ of two variables X₁ and X₂ acting together.

	a. interactions
	b. multicollinearity
	c. residuals
	d. autocorrelation

.

Fill in the blank. In multiple regression analysis, ________ occurs when independent predictors are correlated with one another.

	a. heteroscedasticity
	b. multicollinearity
	c. elasticity
	d. homoscedasticity

.

Fill in the blank. Multicollinearity occurs when ________.

	a. two or more predictor variables in a multiregression are correlated
	b. the probability distribution for the response variable has the same standard deviation regardless of the value of the predictors
	c. the values of a variable at different points in time are correlated with itself
	d. None of the above

.

Fill in the blank. ______ is used to determine whether an additional variable makes a significant contribution to a multiple regression model.

	a. An F test
	b. A Z test
	c. A t test
	d. A chi-square test

.

For a multiple regression problem with four predictors, how many possible models are there?

	a. 25
	b. 16
	c. 15
	d. 10

.

In multiple regression, why do you want to calculate partial correlation?

	a. To investigate the relationship between two variables, controlled for the effects of other variables.
	b. To investigate the relationship between two variables within a portion of the sample.
	c. To remove independent variables from the regression.
	d. To find out which independent variables are the most predictive.

.

R_j² is the coefficient of determination of the model that includes all predictors except the jth predictor. For x₁, R₁² = 0.95, calculate the variance inflation factor for x₁ and determine whether there is a problem with multicollinearity.

	a. 1.05 and there is no problem with multicollinearity
	b. 25 and there is no problem with multicollinearity
	c. 20 and there is a problem with multicollinearity
	d. 10 and there is a problem with multicollinearity

.

R_j² is the coefficient of determination of the model that includes all predictors except the jth predictor. Which of the following formula is used to calculate the variance inflation factor (VIF)?

	a. 1/(1- Rj2)
	b. 1- Rj2
	c. 1+Rj2
	d. 1/(1+ Rj2)

.

What are the consequences of heteroscedasticity in ordinary least square estimates?

	a. Heteroscedasticity results in biased parameter estimates.
	b. Heteroscedasticity leads to inefficient estimation.
	c. Heteroscedasticity leads to inconsistency.
	d. All of the above.

.

Which of the following assumptions about the error term of multiple regression model is NOT true?

	a. Constant variation
	b. Independent
	c. A mean of zero
	d. Exponentially distributed

.

Which of the following criteria is used to determine the best regression model?

	a. The best model minimizes prediction errors.
	b. The best model should be as simple as possible, with the least number of independent variables.
	c. The best model only includes predictors that make a significant contribution to the model.
	d. All of the above.

.

Which of the following is NOT true about multicollinearity?

	a. Multicollinearity is a result of strong correlations between independent variables.
	b. Multicollinearity reduces the variances of the parameter estimates.
	c. Multicollinearity can be reduced by combining the involved variables into one.
	d. Multicollinearity can lead to wrong signs and magnitudes of regression coefficient estimates.

.

Which of the following procedures is backward elimination procedure for selecting models?

	a. The starting model is the one with all the predictors in it, and at each step the procedure tries to drop out one nonsignificant predictor, stopping when all predictors are significant.
	b. The procedure starts without any predictors and tries to add them, step by step, while it also tries to drop out predictors at each step.
	c. The procedure begins with the model having no predictors at all and adds the best available predictor at each step.
	d. None of the above.

.

Which of the following statements describes “redundancy” in multiple regression?

	a. Redundancy occurs when two or more independent variables convey approximately the same predictive information about the dependent variable, consequently, the model using these predictors has predictive power similar to those models using only one of the predictors.
	b. Redundancy occurs when the number of predictors is greater than ten.
	c. Redundancy occurs when adding another variable does not increase predictive power of the regression.
	d. None of the above

.

You are tasked with predicting the heights of sons, which are influenced by genes from both sides of the family. You already have data on the heights of their mothers. You are now allowed to measure a second predictor variable. Choose among the following predictors a second variable that would help most with the predictions.

	a. The height of an uncle on the mother’s side
	b. The height of an aunt on the father’s side
	c. The height of an in-law
	d. The height of a friend on the father’s side

.

Determine the coefficients of an exponential regression y = α.β^x for the following data: x = 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10 and y = 63, 76, 92, 105.7, 122.8, 131.7, 151.3, 179.3, 203.3, 226.54, 248.7, 281.4.

	a. α = 4.3 and β = 0.12
	b. α = 71.5 and β = 1.13
	c. α = 12.5 and β = 10.2
	d. α = 11.5 and β = 0.13

.

In a logistic regression, what is a logit?

	a. The natural logarithm of the odds ratio
	b. The natural logarithm of the probability
	c. The natural logarithm of the risk factors
	d. None of the above

.

In a period of eight years, the revenue of a company (in $ millions) was 25, 26, 28, 27, 26, 29, 29, and 34. For the same period, the spending on advertisement (in $ thousands) was 16, 13, 15, 14, 16, 17, 19, and 22. Estimate the percentage of variation in revenue explained by variation in advertising.

	a. 71%
	b. 10%
	c. 25%
	d. 41%

.

In logistic regression, which test is used to reject the null hypothesis that β = 0?

	a. t-test
	b. F-test
	c. Wald test
	d. Mann-Whitney test

.

In the context of Generalized Linear Models (GLM), what is the link function for logistic regression?

	a. g(p) = p
	b. g(p) = ln(p)
	c. g(p) = p-1
	d. g(p) = ln(p/(1-p))

.

In the context of Generalized Linear Models, which of the following statements is NOT true?

	a. The link function provides the relationship between the linear predictor and the mean of the distribution function.
	b. Each outcome of the dependent variables is assumed to be generated from a distribution in the exponential family.
	c. Generalized Linear Models and General Linear Models refer to the same type of model.
	d. Generalized Linear Models include logistic regression, exponential regression, and multiple linear regression.

.

In the model y = a₀ + a₁x + a₂x², what is the value of a₂ if the relationship between y and x is linear?

	a. 1
	b. a0
	c. 0
	d. –( a0 + a1)

.

In the model y = a₀ + a₁x + a₂x², what is the value of x at which the mean of y takes its maximum if the parabola is mound-shaped or its minimum if the parabola is bowl-shaped?

	a. –a1/2a2
	b. a1/a2
	c. 0
	d. a1/2a2

.

What does heteroscedasticity mean?

	a. The mean of the errors is not zero.
	b. The variances of the predictors are not constant.
	c. The variance of the dependent variable is small.
	d. The variance of the errors is not constant.

.

What is the principle of parsimony when building models?

	a. Models should have no more parameters than necessary to represent the relationship adequately.
	b. All models are wrong; some are useful.
	c. The best regression models are linear.
	d. When building models, one should try to use as many predictors as possible.

.

What is the reason for using polynomial regression?

	a. Polynomial regression is used to address multicollinearity.
	b. The shape of data is not always linear and an nth polynomial can be used to create a better fit.
	c. When your dependent variable is categorical, polynomial regression will provide a better fit.
	d. None of the above.

.

Which of the following is exponential regression model?

	a. y = a0 + a1x + a2x2+…..amxm +ε
	b. y = a0 + a1x1 +ε
	c. y = log(a0 + a1x1)+ε
	d. y = a0 + a1a2x +ε

.

Which of the following is polynomial regression model?

	a. y = a0 + a1x + a2x2+…..amxm +ε
	b. y = a0 + a1x1 +ε
	c. y = log(a0 + a1x1)+ε
	d. y = a0 + a1a2x +ε

.

Which of the following is the equation for logistic regression?

	a. p(x) = β0 + β ∙x
	b. log(p(x)/(1-p(x))) = β0 + β ∙x
	c. log(p(x))= β0 + β ∙x
	d. p(x)/(1-p(x) = β0 + β ∙x

.

Which of the following is Type I error?

	a. Type 1 error is made when a test fails to reject a false null hypothesis.
	b. Type 1 error is made when a test rejects a true null hypothesis.
	c. Type 1 error is made when a test rejects a false null hypothesis.
	d. None of the above.

.

Which of the following method is used in data mining?

	a. Logistic regression
	b. Support vector machine
	c. Naïve Bayse classification
	d. All of the above

.

In an epidemiological study focusing on a biomarker A, blood samples were collected from two different ethnic groups, “Caucasian” and “Asian.” For the Caucasian group, the measured values of biomarker A in seven individuals are 42, 41, 47, 42, 37, 41, and 35. For the Asian group, the values of biomarker A in seven individuals are 16, 31, 20, 35, 47, 27, and 31. The null hypothesis is that the probability distributions of biomarker A in Caucasian and Asian populations are identical. You are asked to perform a Wilcoxon rank-sum test to reject or confirm the null hypothesis. What are the sums of the ranks for observations in the Caucasian and Asian groups, respectively?

	a. 30 and 50
	b. 80 and 10
	c. 60 and 30
	d. 70 and 35

.

Which non-parametric test would you use to assess whether there is a significant difference between the mean ranks of two conditions?

	a. Wilcoxon
	b. Kruskal-Wallis
	c. F-test
	d. None of the above

.

Which nonparametric test is equivalent to a parametric t-test for independent sample means?

	a. Wilcoxon signed-rank test
	b. Mann-Whitney test
	c. Kruskal-Wallis test
	d. None of the above

.

Which of the following is a method of multivariate analysis?

	a. Principal Component Analysis
	b. Multiple regression
	c. Hypothesis testing
	d. All of the above

.

Which of the following is not a nonparametric test?

	a. Sign test
	b. Wilcoxon sign-rank test
	c. Kolmogorov–Smirnov test
	d. Mann-Whitney-Wilcoxon test

.

Which of the following is NOT a parametric procedure?

	a. Pearson’s correlation coefficient
	b. ANOVA
	c. A sign test
	d. Logistic regression

.

Which of the following is NOT true about nonparametric tests?

	a. Nonparametric tests are more likely to reject the null hypothesis, compared with their parametric equivalents.
	b. Nonparametric tests use the Z-distribution for large samples.
	c. Nonparametric tests are cumbersome to use for large sample sizes.
	d. Nonparametric tests are more conservative as compared with their parametric equivalents.

.

Which of the following statement is correct?

	a. All variables in a multiple regression analysis must be quantitative.
	b. All variables in a multiple regression analysis must be positive.
	c. All variables in a multiple regression analysis must be qualitative.
	d. None of the above.

.

Which of the following statements about nonparametric tests is NOT true?

	a. If the sampled populations are normally distributed, parametric tests such as F and t tests are more powerful than their nonparametric counterparts.
	b. The Wilcoxon rank-sum test requires that we take independent random samples.
	c. Nonparametric tests can be easily extended to multiple regression.
	d. The Kruskal-Wallis test is the nonparametric counterpart of the one-way ANOVA test.

.

Which test can be used to test normality of data?

	a. t-test
	b. Kolmogorov–Smirnov test
	c. Mann-Whitney test
	d. F-test

.

A group of shapes is classified into five circles and nine rectangles. Determine the expected information (entropy) needed to classify a tuple in the group.

	a. 0.14 bits
	b. 0.74 bits
	c. 0.94 bits
	d. 0.54 bits

.

An event happens at a constant hazard rate of 0.5 events per second. What is the survival function at 1 second since the last event?

	a. 0.61
	b. 0.39
	c. 0.41
	d. 0.59

.

D has fourteen shapes, classified into five circles and nine rectangles. Determine the Gini coefficient for D.

	a. 0.22
	b. 1.50
	c. 0.92
	d. 0.46

.

Fill in the blank. In data mining, ______ is used to predict categorical class labels.

	a. Classification
	b. Numerical prediction
	c. Hypothesis testing
	d. None of the above

.

Fill in the blank. In decision tree induction, ______ is used to measure impurity.

	a. Information gain
	b. Gini coefficient
	c. Entropy
	d. Number of nodes

.

Fill in the blank. In naïve Bayes’ classification, the Bayes’ theorem can be informally written as ______.

	a. posteriori = likelihood x prior / evidence
	b. prior = likelihood x posteriori / evidence
	c. posteriori = evidence x prior / likelihood
	d. posteriori = likelihood x evidence/ prior

.

In principal component analysis (PCA), what does “reduction of dimensionality” mean?

	a. PCA reduces a large number of interrelated variables to a small number of uncorrelated principal components.
	b. PCA linearizes the problem.
	c. PCA reduces the variance of variables.
	d. PCA identifies the number of independent variables in the original problem.

.

In structural equation modeling (SEM), what are endogenous variables?

	a. Explanatory variables, only appearing in the structural equations
	b. Errors in the structural equations
	c. Response variables whose values are determined by the model
	d. Variables uncorrelated with the error of the structural equations

.

In support vector machine, which term is used to refer to the separation between classes?

	a. Margin
	b. Hyperplane
	c. Support vectors
	d. None of the above

.

In survival analysis, if the probability of surviving to time t is P(t), what is the survival function?

	a. P(t)
	b. 1-P(t)
	c. 1+P(t)
	d. P(t)2

.

Which component of a time series refers to the long-run decline or growth in a time series?

	a. trend
	b. seasonal
	c. irregular
	d. cyclical

.

Which of the following activities is considered to be part of data mining?

	a. Classification
	b. Outlier analysis
	c. Clustering
	d. All of the above

.

Which of the following approaches to modeling time series is the Box-Jenkins approach?

	a. The approach is based on a linear regression of the current value of the series against the white noise of one or more prior values of the series.
	b. Frequency domain is used to analyze the time series.
	c. The time series is decomposed into trend, seasonal, and residual components.
	d. The approach combines moving average and autoregressive approaches.

.

Which of the following is NOT true about exponential smoothing in time-series analysis?

	a. Recent observations are given relatively more weight in forecasting than the older observations.
	b. The weights assigned to the observations are equal to 1/N, where N is the number of observations.
	c. In single exponential smoothing, the weights decrease geometrically.
	d. Single exponential smoothing does follow the data well when there is a trend.

.

Which of the following is NOT true about factor loadings in factor analysis?

	a. Factor loadings are the correlation coefficients between the variables and the factors.
	b. Factor loading of a variable is the percentage of variable in that variable explained by the factor.
	c. In practice, factor loadings should be 0.95 or higher to confirm that independent variables are represented by a particular factor.
	d. If the communality (the sum of the squared factor loadings for all factors for a given variable) exceeds 1.0, there is a spurious solution.

.

Which of the following is NOT true about hazard function?

	a. A hazard function can be interpreted as the expected number of events per individual per unit of time.
	b. For an exponential survival function, the hazard rate is a constant.
	c. The probability of an event is the cumulative hazard function.
	d. The hazard function can be derived from the survival function.

.

Which of the following is NOT true about principal component analysis (PCA)?

	a. PCA transforms data to a new coordinate system, with each coordinate being referred to as a principal component.
	b. In PCA, each of the principal components is a nonlinear combination of the original variables.
	c. The principal components are arranged in order of decreasing variance.
	d. The most informative principal component is the first component.

.

Which of the following is not true about structural equation models (SEM)?

	a. Structural-equation models are multiple-equation regression models in which the response variable in one regression equation might be an explanatory variable in another equation.
	b. Path diagrams can be used to represent an SEM in the form of a causal graph.
	c. In a recursive SEM, causation in the model is bidirectional.
	d. SEMs can include variables that are indirectly measured through their effects.

.

Which of the following is NOT true about supervised learning?

	a. Observations or measurements are labeled with predefined classes.
	b. In supervised learning, class labels of the data are not known.
	c. Supervised learning consists of two steps: training and testing.
	d. Decision tree induction is a supervised learning algorithm.

.

Which of the following is true about endogenous and exogenous variables in structural equation models (SEMs)?

	a. There is one structural equation for each endogenous variable.
	b. An endogenous variable may appear as an explanatory variable in other structural equations.
	c. Exogenous variables are determined outside of the model.
	d. All of the above.

.

Which of the following is true about the Cox proportional hazard model?

	a. The Cox proportional hazard model is a parametric model.
	b. The Cox proportional hazard model is a semi-parametric model.
	c. The Cox proportional hazard model cannot accommodate time-dependent covariates.
	d. All of the above.

.

Which of the following is(are) potential problem(s) for a Cox regression model in survival analysis?

	a. Violation of the assumption of proportional hazards
	b. Influential observations
	c. Nonlinearity in the relationship between the covariates and the log-hazard
	d. All of the above

.

Which of the following software package is an implementation of support vector machine in R?

	a. e1071
	b. lm
	c. arima
	d. None of the above

.