- | P a g e
ISYE 6501 Midterm 1 Exam Newest 2025 Complete 100 questions and Correct Detailed Answers (Verified Answers) |Already Graded A+
A survey of 25 people recorded each person's family size and type of car. Which of these is a data point?
- The 14th person's family size and car type
- The 14th person's family size
- The car type of each person - ANSWER-A.
observation/record of measurements for a single member of a population or data set. In the standard tabular format, a row of data.
The farther the wrongly classified point is from the line...... - ANSWER-The bigger the mistake we've made
The term including the margin gets larger so the importance of a large margin out weights avoiding mistakes and classifying known data samples - ANSWER-As lambda gets larger
Each column is information about what? - ANSWER-An observation 1 / 4
- | P a g e
What is a Support Vector Macine? - ANSWER-Basic Machine Learning; classifying a binary response based on a set of attributes and data points.
In SVM, do we scale the data? - ANSWER-Must scale the data!
In SVM, what is C? - ANSWER-C Value or Cost - this is our tradeoff factor between correctness and margin
What is Box Cox Transformation? - ANSWER-useful for transforming a response to eliminate heteroscedasticity; logarithmic transformation; stretches out the smaller range to enlarge its variability; shrinks the larger range to reduce its variability. The idea is to find the best value of lambda; first check the Q-Q plot to see if you need to do it.
What is hteroscedasticity? - ANSWER-Unequal variance; might cause bias in our data b/c variance at the upper end can make estimation errors larger and push the model to fit those data points better
- / 4
- | P a g e
When should detrending be considered? - ANSWER-Response, Predictors, Factor-based model, regression, SVM, etc. to analyze time series data
How to detrend - ANSWER-Factor by Factor; i.e.: simple linear
regression Price = 45,600 +23.2 * Year Detrend Price = Actual Price - (45,600 +23.2*Year)
What is commong scaling - ANSWER-data between 0 and 1 which is linearly
what is standard scaling - ANSWER-scaling to normal
distribution: mean of 0, std = 1; standardization
when to use type of scaling method? - ANSWER-Data in a
bounded range is good for scaling like: neural networks, SAT
score; Data sometimes in standardization like PCA and clustering;
- / 4
- | P a g e
K-Nearest Neighbor Algorithm - basic machine learning -
ANSWER-Find the class of a point: pick the k closest points to
the new one The new point's class is the most common among the k neighbors
For K-Nearest Neighbor keep in mind: - ANSWER-1. Can use
other distance metrics;
- attributes can be weighted by importance
- unimportant attributes can be removed.
- Choose a good value of k
How do we evaluate the model output for SVM? - ANSWER-1.look at coefficients of the predictors, if close to zero they are probably not good.
- C Value - up or down to trade off error rate and margin
- Accuracy - data point accurately
CUSUM - ANSWER-short for cumulative sum; change detection method that compares observed distribution mean with a threshold level of change. C and T control values.
- / 4