Question 1

This technique associates a conditional probability value with each data instance.&#10;A) linear regression&#10;B) logistic regression&#10;C) simple regression&#10;D) multiple linear regression

Accepted Answer

Logistic regression is a technique that models the probability of a binary response variable based on one or more predictor variables. It provides conditional probabilities for each data instance, which makes it suitable for classification tasks. Linear regression, simple regression, and multiple linear regression are techniques used for modeling continuous response variables and do not provide conditional probabilities.

Question 2

Logistic regression is a ________ regression technique that is used to model data having a _____outcome.&#10;A) linear, numeric&#10;B) linear, binary&#10;C) nonlinear, numeric&#10;D) nonlinear, binary

Accepted Answer

Logistic regression is a nonlinear regression technique that is used to model data having a binary outcome, such as yes or no, true or false, and success or failure. It estimates the probability of an event occurring based on one or more predictor variables, and it is commonly used in data analysis and machine learning for classification tasks. Other regression techniques, such as linear regression, are used for modeling data with numeric outcomes.

Question 3

The leaf nodes of a model tree are&#10;A) averages of numeric output attribute values.&#10;B) nonlinear regression equations.&#10;C) linear regression equations.&#10;D) sums of numeric output attribute values.

Accepted Answer

The leaf nodes of a model tree are linear regression equations, which are used to predict a numeric value for each instance that reaches the leaf. This approach allows for both the segmentation of the data (through the tree structure) and the application of a simple linear model to each segment for prediction purposes.

Question 4

The probability of a hypothesis before the presentation of evidence.&#10;A) a priori&#10;B) subjective&#10;C) posterior&#10;D) conditional

Accepted Answer

The term "a priori" means "before the fact" and refers to the probability of a hypothesis before any evidence is presented. It is based on prior knowledge, experience, and beliefs about the likelihood of the hypothesis being true.

Question 5

This clustering algorithm merges and splits nodes to help modify nonoptimal partitions.&#10;A) agglomerative clustering&#10;B) expectation maximization&#10;C) conceptual clustering&#10;D) K-Means clustering

Accepted Answer

The answer of This clustering algorithm merges and splits nodes...

Question 6

This supervised learning technique can process both numeric and categorical input attributes.&#10;A) linear regression&#10;B) Bayes classifier&#10;C) logistic regression&#10;D) backpropagation learning

Accepted Answer

The Bayes classifier is able to process both numeric and categorical input attributes through probability calculations and classification rules. Linear regression and logistic regression models are primarily used for numerical input attributes, while backpropagation learning is a technique used in artificial neural networks for various types of data including numerical, categorical, and even image data.

Question 7

Simple regression assumes a __________ relationship between the input attribute and output attribute.&#10;A) linear&#10;B) quadratic&#10;C) reciprocal&#10;D) inverse

Accepted Answer

Simple regression assumes a linear relationship between the input attribute and output attribute, meaning that the change in the input attribute is directly proportional to the change in the output attribute. The other options (B, C, and D) represent non-linear relationships.

Question 8

Regression trees are often used to model _______ data.&#10;A) linear&#10;B) nonlinear&#10;C) categorical&#10;D) symmetrical

Accepted Answer

Regression trees are particularly useful for nonlinear data as they can capture complex nonlinear relationships between variables. For linear data, a simple linear regression model may be more appropriate. Categorical data may require different types of models such as logistic regression or decision trees. Symmetry is not a characteristic of the data that would inform the choice of modeling technique.

Question 9

Machine learning techniques differ from statistical techniques in that machine learning methods&#10;A) typically assume an underlying distribution for the data.&#10;B) are better able to deal with missing and noisy data.&#10;C) are not able to explain their behavior.&#10;D) have trouble with large-sized datasets.

Accepted Answer

Machine learning methods are designed to handle missing, noisy, and complex data sets, while statistical techniques rely on having complete and accurate data. Machine learning also focuses on predicting outcomes, whereas statistical analysis is more concerned with understanding the relationship between variables. However, machine learning models often lack interpretability, meaning it can be difficult to understand why the model made a particular prediction. Additionally, machine learning can be computationally intensive and may struggle with very large datasets.

Question 10

This unsupervised clustering algorithm terminates when mean values computed for the current iteration of the algorithm are identical to the computed mean values for the previous iteration.&#10;A) agglomerative clustering&#10;B) conceptual clustering&#10;C) K-Means clustering&#10;D) expectation maximization

Accepted Answer

The described termination condition is specific to the K-Means clustering algorithm, which iteratively assigns data points to the nearest cluster center, updates the mean values for each cluster, and repeats until convergence. Once the means for the current iteration are identical to the means for the previous iteration, further iterations will not change the cluster assignments, so the algorithm terminates. Agglomerative clustering and conceptual clustering do not use mean values in this way, and EM does not necessarily converge when mean values are identical between iterations.

Question 11

This clustering algorithm initially assumes that each data instance represents a single cluster.&#10;A) agglomerative clustering&#10;B) conceptual clustering&#10;C) K-Means clustering&#10;D) expectation maximization

Accepted Answer

The answer of This clustering algorithm initially assumes that each...

Question 12

With Bayes classifier, missing data items are&#10;A) treated as equal compares.&#10;B) treated as unequal compares.&#10;C) replaced with a default value.&#10;D) ignored.

Accepted Answer

The answer of With Bayes classifier, missing data items are&#10;A)...

Question 13

The table below contains counts and ratios for a set of data instances to be used for supervised Bayesian learning. The output attribute is sex with possible values male and female. Consider an individual who has said no to the life insurance promotion, yes to the magazine promotion, yes to the watch promotion and has credit card insurance. Use the values in the table together with Bayes classifier to determine which of a,b,c or d represents the probability that this individual is male.

A) 4/6) 2/6) 2/6) 2/6) 6/10) / PE)
B) 4/6) 2/6) 3/4) 2/6) 3/4) / PE)
C) 4/6) 2/6) 4/6) 2/6) 6/10) / PE)
D) 2/6) 4/6) 4/6) 2/6) 4/10) / PE)

Accepted Answer

The answer of The table below contains counts and ratios...

Deck 10: Statistical Techniques