WorksheetsLesson 3 - Model Training
Total questions: 20
Worksheet time: 10mins
Which of the following are keys to model training?
Data Preparation
Data Management
Data Handling
Data Diversion
Multiple machine learning models can together improve the results. What kind of learning is this?
Collective Learning
Symbiotic Learning
Ensemble Learning
More Learning
Who contains resources for exploring, transforming, and managing data in Azure ML.
Datasets
Datastores
The data management process, includes...
The use of datastores and datasets
Data importing
Versioning
How to monitor for data drift
Which of the following could not be a feature of a T-Shirt
Medium
Colour
Size
None of the above
Which of these could be steps of feature engineering?
Derive new features on the basis of existing features
Applying a mathematical function to a pre-existing feature
Training a separate ML model to create new features
Fetching more data with wider set of features
Which dataset is supported in Microsoft Azure?
Tabular Dataset
Web URL (File Dataset)
Graph Based Data set
Folder Based Data set
Data Wrangling is ___________ process.
Done according to performance, An iterative
Done only once, non-iterative
Datasets helps in seamless integration with numerous machine learning task, such as:
Consummation of Datastores in automated processes
Data labelling
Monitoring issues like data drift
Access of datastore through inference process
Identify the correct sequence.
Dataset Monitor -> Datastore -> Dataset
Datastore -> Dataset Monitor -> Dataset
Data Store -> Dataset -> Dataset Monitor
In which technique in ML is use to decrease number of features
Feature deduction
Feature subtraction
Feature reduction
Deep learning depends on feature engineering much more than Classical machine learning.
True
False
what's the potential benefits of feature engineering
Smaller trained model
Increased model accuracy
Smaller space taken by model
Faster training time
Driving a Boolean value from each entity is called
Part-of
Binning
Flagging
Aggregation
A change in user behavior over time is
Data quality issue
Covariate shift
Natural drift in data
Upstream process change
Two features that used to be correlated are no longer correlated is called
Data quality issue
Covariate shift
Natural drift in data
Upstream process change
Missing values imputation comes under
Data Cleansing
Data restructuring
Data discovery
Data exploration
Data used to check the performance of the final, fully trained model.
Train data
Test data
Hyper parameters
Validation data
The centralized place for working with all the components of the machine learning process
Model registry
Experiments
Workspace
Compute Instance
Classify an image as one (and only one) of five possible types of birds.
Multi-class single label classification
Multi-class multi label classification
Binary classification
