What are the main challenges in training a Machine Learning model with imbalanced datasets?
- Computational complexity
- Dimensionality reduction
- Lack of suitable algorithms
- Overfitting to the majority class
Training on imbalanced datasets can lead to models that are biased towards the majority class, since they have seen more examples of it. This can make the model perform poorly on the minority class.
While estimating the coefficients in Simple Linear Regression, you find that one of the assumptions is not met. How would this affect the reliability of the predictions?
- Increase Accuracy
- Make Predictions More Reliable
- Make Predictions Unreliable
- No Effect
If the assumptions of Simple Linear Regression are not met, the reliability of the predictions may be compromised, and the model may become biased or inefficient.
You have a dataset with many correlated features, and you decide to use PCA. How would you determine which Eigenvectors to keep?
- By choosing the eigenvectors with the highest eigenvalues
- By randomly selecting eigenvectors
- By selecting the eigenvectors with negative eigenvalues
- By using all eigenvectors without exception
You would keep the eigenvectors corresponding to the highest eigenvalues, as they explain the most variance in the data. The lower the eigenvalue, the less significant the corresponding eigenvector.
Explain the application of clustering algorithms in customer segmentation for marketing strategies.
- Clustering Customers
- Image Recognition
- Supply Chain Management
- Text Classification
Clustering algorithms are used in customer segmentation to group customers based on similar characteristics or behaviors. These clusters help marketing teams to target specific segments with tailored marketing strategies, improving engagement and conversion rates.
What are the underlying assumptions of Logistic Regression?
- Linearity of predictors and log-odds, Independence of errors, No multicollinearity
- Linearity, Independence, Normality, Equal Variance
- No assumptions required
- Nonlinearity, Dependence, Non-Normality
Logistic Regression assumes a linear relationship between predictors and log-odds, independence of errors, and no multicollinearity among predictors. It does not assume normality or equal variance of errors.
A dataset contains both categorical and numerical features. Which ensemble method might be suitable, and what preprocessing might be required?
- Random Forest with no preprocessing
- Random Forest with normalization
- Random Forest with one-hot encoding
- Random Forest with scaling
Random Forest is an ensemble method suitable for handling both categorical and numerical features. For categorical features, one-hot encoding might be required to convert them into a numerical format that the algorithm can process.
Describe a scenario where Hierarchical Clustering would be more beneficial than K-Means Clustering, and explain the considerations in choosing the linkage method.
- When a fixed number of clusters is required
- When clusters are uniformly distributed
- When clusters have varying sizes and non-spherical shapes
- When computational efficiency is the priority
Hierarchical Clustering is more beneficial than K-Means when clusters have varying sizes and non-spherical shapes. Unlike K-Means, Hierarchical Clustering does not assume spherical clusters and can handle complex structures. The choice of linkage method will depend on the specific characteristics of the clusters, with considerations like distance metric and desired cluster shape guiding the selection.
In reinforcement learning, the agent learns to take actions that maximize the cumulative __________.
- accuracy
- errors
- loss
- rewards
In reinforcement learning, the agent tries to maximize cumulative rewards through its actions.
Machine Learning is commonly used in ____________ to create personalized recommendations.
- Drug Development
- Recommender Systems
- Traffic Management
- Weather Prediction
Machine Learning is extensively used in Recommender Systems to create personalized recommendations, analyzing user behavior and preferences.
What are the underlying assumptions of LDA, and how do they affect the performance?
- Assumes different covariance matrices, normal distribution; affects adaptability
- Assumes equal class sizes; affects bias
- Assumes equal variance, non-normal distribution; affects robustness
- Assumes normal distribution, equal covariance matrices; affects classification accuracy
LDA assumes that the features are normally distributed and that the classes have equal covariance matrices. These assumptions, if met, lead to better "classification accuracy," but if violated, may lead to suboptimal performance.