ML Machine Learning Flashcards
Certification Study
What distinguishes the ML process?
Data fed model with labeled data.
What is labeled data?
Data is labeled or by a by a targeted variable or output or classification either provided by humans or through a reliable process.
How is Unlabeled data distinguished?
The instances or examples do not have any associated labels or target variables. There is ONLY input features.
What are the main types of data used for training.
- Structured data
- Unstructured data
What is structured data?
Data is organized and formatted in a predefined way such as tables or databases with rows and columns.
When is it best to use structured data?
Use structured data for traditional learning algorithms that require defined features and labels.
Provide some examples of structured data.
- Tabular data: Data is stored in spreadsheets, databases or CSV files where rows represent instances and columns represent features or attributes.
2.Time Series Data: Data consists of sequences of values measure at successive points in time like stick market process, sensors or weather data.
Explain what is unstructured data and provide examples of each of the 2 different types.
Unstructured data does not have a predefined structure like text, images and video. More advance learning techniques are needed to extract meaningful patterns and insights.
Text data: documents, articles, social media posts and other textual data
Image data: Includes digital images, photographs, video frames
How is the ML learning process? Are there defined categories and if so list them.
Data is fed into ML learning algorithms and divided into 3 categories.
1. Supervised Learning
2. Unsupervised Learning
Reinforcement Learning
Briefly describe Supervised learning.
In Supervised Learning Algorithms are trained on LABELED DATA.
The goal is to learn a mapping function that can predict the output for new, unseen inout data.
Briefly describe UnSupervised Learning.
In UnSupervised learning algorithms learn from UNLABELED DATA.
The goal is discover inherent patterns, structures or relationships in the input data.
Briefly describe Reinforcement Learning.
In Reinforcement learning: the machine is rewarded with a performance score in a semi supervised learning with only a portion of the data is labeled.
Feedback is rewards or punishments to improve decision making.
What is batch inferencing and when do you utilize it?
Batch inferencing is used when the computer takes in alot of data - images, text and needs to analyze it all at once.
Used for ACCURACY NOT SPPED.
When do you use REAL-Time inferencing and what is is actually?
Needed for speed and quick response such as chatbots or self-driving cars. Real time or instant analysis is needed.
After a model is trained it can be used in a process to make predictions or decisions based - what is this process called? Is there more than one type? - if so list it or them.
The process a trained ML uses is called inferencing.
2 types
1. Batch Inferencing
2. Real-time inferencing
