Question 1

A data analyst has performed a cluster analysis on a dataset and generated the following cluster matrix:

Cluster 1 Cluster 2

Count 150 200

Mean X 5.0 2.0

Mean Y 2.0 5.0

SSD 20.0 15.0

Given the cluster matrix above, which statement correctly interprets the cluster characteristics?

Options :

A :

Cluster 1 has a larger variability within the cluster than Cluster 2, as indicated by the sum of squares distance (SSD).

B :

Cluster 1 is more densely packed than Cluster 2 because it has a higher count.

C :

The two clusters are likely well separated along the Y dimension given the difference in their mean Y values.

D :

The two clusters are likely well separated along the X dimension given the difference in their mean Y values.

Answer: C

Question 2

You are working on a predictive model and you notice that your categorical variable "Color" has 50 different levels, which is adding complexity to your model. You decide to group these levels into fewer categories based on their frequency of occurrence to improve the model’s interpretability and performance. What technique would you most likely use for this task?

Options :

A :

Principal Component Analysis (PCA)

B :

Decision Tree Binning

C :

Top-N Binning

D :

One-hot Encoding

Answer: C

Question 3

When using SAS Visual Statistics to build a generalized linear model for count data, which model settings should you choose to properly account for overdispersion in the data if the initial model assessment indicates that the variance is greater than the mean?

Options :

A :

Use a normal distribution with an identity link function and adjust the model for overdispersion.

B :

Use a Poisson distribution with a log link function, and apply a scale parameter to adjust for overdispersion.

C :

Use a negative binomial distribution with a log link function.

D :

Use a Poisson distribution with a square root link function and adjust the model for overdispersion.

Answer: C

Question 4

You have built several predictive models using training data and now you want to assess the models' performance using validation data. Which measure is NOT appropriate for comparing model performance in terms of model bias for continuous outcomes?

Options :

A :

Root Mean Squared Error (RMSE)

B :

Akaike’s Information Criterion (AIC)

C :

Mean Absolute Error (MAE)

D :

Coefficient of Determination (R-squared)

Answer: B

Question 5

When preparing data for a predictive modeling project, a data scientist notices that the categorical variable 'payment_type' with four categories ('credit card', 'debit card', 'paypal', 'other') exhibits a high degree of variability in the outcome variable (purchase amount). To improve the model's predictive accuracy, what strategy can the data scientist use to handle the 'payment_type' variable?

Options :

A :

Group all categories into a single 'payment_type' indicator variable.

B :

Perform one-hot encoding on 'payment_type' to create a binary indicator for each category.

C :

Collapse 'paypal' and 'other' into a single category if they have similar effects on the purchase amount.

D :

Treat 'payment_type' as a continuous variable to simplify the model.

Answer: B