Podcast
Questions and Answers
What is the aim of cluster analysis?
What is the aim of cluster analysis?
- To color various kinds of graphs according to selected properties
- To calculate simple statistics for large data sets
- To divide a group of objects into clusters so that the objects within a cluster are similar (correct)
- To draw various kinds of graphs for data visualization
What are the key steps involved in cluster-based compound selection?
What are the key steps involved in cluster-based compound selection?
- Calculate simple statistics, use a clustering algorithm, divide molecules into subsets, select a representative subset
- Draw various kinds of graphs, color according to selected properties, calculate simple statistics, divide molecules into subsets
- Generate descriptors for each compound, calculate similarity, use a clustering algorithm, select a representative subset (correct)
- Generate descriptors for each compound, use a clustering algorithm, color according to selected properties, select a representative subset
What is the purpose of dividing molecules into subsets in HTS data sets?
What is the purpose of dividing molecules into subsets in HTS data sets?
- To calculate simple statistics for large data sets
- To color various kinds of graphs according to selected properties
- To draw various kinds of graphs for data visualization
- To help navigation through the data (correct)
What is the graphical representation used for the property distributions of active, moderately active, and inactive compounds?
What is the graphical representation used for the property distributions of active, moderately active, and inactive compounds?
What does a representative object chosen from each cluster indicate in the context of compound selection?
What does a representative object chosen from each cluster indicate in the context of compound selection?
What is used to visualize the relationships between clusters in hierarchical clustering?
What is used to visualize the relationships between clusters in hierarchical clustering?
What is the main characteristic of a linear discriminant analysis in drug classification?
What is the main characteristic of a linear discriminant analysis in drug classification?
How is the appropriate number of clusters in hierarchical clustering determined?
How is the appropriate number of clusters in hierarchical clustering determined?
What is a key feature of a Feed-Forward Neural Network in drug classification?
What is a key feature of a Feed-Forward Neural Network in drug classification?
What does Multidimensional Scaling involve?
What does Multidimensional Scaling involve?
What does a Kohonen Network do in drug classification?
What does a Kohonen Network do in drug classification?
What is the purpose of Sammon mapping?
What is the purpose of Sammon mapping?
What is a characteristic of Random Forest in drug classification?
What is a characteristic of Random Forest in drug classification?
What is Substructural analysis (SSA) related to?
What is Substructural analysis (SSA) related to?
How do decision trees contribute to drug classification?
How do decision trees contribute to drug classification?
Flashcards are hidden until you start studying
Study Notes
Machine Learning Methods in Drug Classification
- Linear discriminant analysis separates molecules into active and inactive classes, but may not completely separate the data points.
- A linear discriminant analysis uses a discriminant function, a linear combination of independent variables, to compute the activity class for a molecule.
- Feed-Forward Neural Network is a supervised learning method with layers of nodes and connections between them.
- Each node in the Feed-Forward Neural Network exists in a state between 0 and 1, and the network must be trained before use.
- Kohonen Network, an unsupervised learning method, creates regions containing similar nodes based on input data.
- Each node in the Kohonen Network has an associated vector that corresponds to the input data, and the network creates regions containing similar nodes.
- Decision trees provide interpretable rules to associate molecular features and descriptor values with the activity or property of interest.
- Ensemble approaches involve the construction of collections of trees, such as Bagging trees and Random Forest, to classify new molecules using a majority voting mechanism.
- Random Forest is an extension of bagging, where a small subset of descriptors is randomly selected at each node.
- Boosting improves the performance for data points misclassified by its predecessor by giving more weight to such points.
- Decision trees are used to classify unknown molecules by following a path through the tree according to the values of relevant properties.
- Machine learning methods are used in drug classification to predict the activity class for new, unseen molecules, based on their molecular descriptors.
Studying That Suits You
Use AI to generate personalized quizzes and flashcards to suit your learning preferences.
