D491 Introduction to Analytics - Set 5 - Part 2

Correct Answer: Isolation forest

Isolation forest is commonly used for anomaly detection because it isolates outliers by creating random partitions in the data.

Correct Answer: RNNs can capture temporal dependencies in sequential data

RNNs are designed to handle sequential data and capture temporal dependencies, making them suitable for time series forecasting.

Correct Answer: To quantify the difference between predicted and actual class probabilities

Cross-entropy loss measures the difference between predicted and actual probabilities, making it suitable for classification tasks.

Correct Answer: To improve the model's performance by selecting the best hyperparameters

Hyperparameter tuning aims to find the optimal set of hyperparameters that maximize the model's performance on validation data.

Correct Answer: Principal Component Analysis (PCA)

PCA is used to reduce the dimensionality of large datasets by transforming the data into a set of orthogonal principal components.

Correct Answer: To determine the split at each node

Gini impurity measures how often a randomly chosen element would be incorrectly classified, helping to decide where to split in a decision tree.

Correct Answer: Sigmoid

The sigmoid activation function is commonly used in binary classification tasks as it outputs values between 0 and 1, representing probabilities.

Correct Answer: It decreases the computational time

Mini-batch gradient descent updates the model's weights more frequently than full-batch gradient descent, reducing the computational time and allowing for faster convergence.

Correct Answer: Continuous data

Regression analysis is typically performed on continuous data to model the relationship between a dependent variable and one or more independent variables.

Correct Answer: To balance imbalanced datasets

Over-sampling techniques, such as SMOTE, are used to balance imbalanced datasets by increasing the number of instances in the minority class.

Correct Answer: The proportion of predicted positives that are correct

Precision measures the accuracy of positive predictions by comparing the number of true positives to the total number of predicted positives.

Correct Answer: Spectral clustering

Spectral clustering is commonly used for community detection in social network data due to its ability to identify clusters in non-convex shapes.

Correct Answer: To prevent overfitting by randomly dropping neurons during training

Dropout randomly drops neurons during training to prevent the model from overfitting and improve generalization.

Correct Answer: To group similar data points together

Clustering algorithms aim to group similar data points together based on the characteristics of the data, without the need for labeled data.

Correct Answer: Random forest

Random forest is often used for customer churn prediction due to its ability to handle complex data and provide accurate classifications.

Correct Answer: Silhouette score

The silhouette score evaluates the quality of clusters by measuring how close each data point is to its own cluster compared to other clusters.

Correct Answer: Stochastic gradient descent updates the weights after each training example

Stochastic gradient descent updates the weights after each individual training example, while batch gradient descent updates the weights after processing the entire dataset.

Correct Answer: Precision-Recall curve

The precision-recall curve is commonly used to evaluate classification models on imbalanced datasets, as it provides a better understanding of model performance than accuracy.

Correct Answer: Collaborative filtering

Collaborative filtering is widely used in recommendation systems to predict user preferences based on similar users or items.

Correct Answer: To weigh the importance of terms within a document

TF-IDF weighs the importance of words in a document by considering how frequently they appear in the document and across other documents.