Quiz-1 Artificial Intelligence

Which of the following task can be solved using supervised learning

• Sorting the unwanted plastic from garbage

• Categorizing people based on their education level

• Identifying nature of the emoji entered

• Separating vehicles based on various properties

When should we apply machine learning, select all that apply

• The amount of data generated can not be interpreted by hand easily

• When data and results are available

• When we want to increase data

• When data patterns are needed

Which of the following is required about target column for high accuracy.

• Target column should have non-linear relation with predictors

• Target column should have a high correlation value with attributes

• Target column should have linear relation with other predictors

• Target column should be independent of other columns

Which one of the following statement is correct

• All the features should have linear relation with each other

• Features should have high correlation with each other

• Features should be non-linear with each other

• None of the above

Select that are true about un-supervised algorithms

• They are only used for classification of data

• They can be used to separate the data chunks

• They can perform well on categorical data

• They are only used when data is not labeled

Which of the following is not a classification algorithm

• Decision Tree

• Support vector machine

• Linear regression

• Logistic regression

For categorical data with normal distribution and zero outliers which algorithm is best

• Decision Tree

• Linear Regression

• Logistic Regression

• Linear Discriminant Analysis

Which of the following options are beneficial while overfitting

• Tree pruning

• Features reduction

• Increasing data

• Making model complex

Which of the following is the definition of underfitting

• When model is very complex and captures all the patterns as well as noise

• When data is very large and it give high training accuracy

• When model is very weak and it gives low training accuracy

• When model gives high testing accuracy.

What should be done when our model is underfit

• Increase the size of data

• Add more features

• Reduce the complexity of model

• Reduce number of features