Tähän tulee youtube video, sen saa edukamu komponentistä nimeltä testvideo-youtube.yaml
Tähän yhdistä oikeat kohdat- tehtävä
Supervised learning is the easiest type of machine learning to understand and to implement. Supervised learning models can be further separated into two groups based on their approach:
Classification models are able to predict between two or more classes. You will use a classification model when you need to find out in which category does an object belongs. A binary classifier would have only two mutually exclusive classes: [spam, not spam] or [dog, rabbit]. An n-class multiclass classification model will learn to predict n amount of classes: [1,2,3,4,5,6,7] or [background, dog, cat, person, ball].
Regression models are able to predict continuous-valued attributes. The output is often a float-point number such as 12.53255. The attribute which is being predicted could be, for example, life expectancy in years, housing price in thousands of dollars, annual profit of a company in dollars.
In the regression example, the predicted value is the house price in dollars. The x-axis would be some feature of the house, for example, the area in square meters, and the y-axis would be the price. In the classification example, we are classifying dogs and rabbits. Both the x and y axes would be different features of the animals.
In unsupervised learning, you will not rely on learning from labeled data. Instead, the algorithm is data-driven and will extract insights based on the dataset.
Unsupervised learning applications include clustering data, anomaly detection, and dimensionality reduction. Clustering data may be useful in any kind of application, such as in market segmentation.
Reinforcement learning systems learn by interacting with the environment. The action will change the current state of the environment and produce some reward. Learning happens as a "loop of trial and error".
A number of iterations later, the system has learned to maximize the long-term reward by taking actions based on the current state of its environment. Reinforcement learning is useful in dynamic environments, such as in games and automated stock market trading.
Read through the attached material and do tasks 1 and 2. Read pages 1-14 from the book A Brief Introduction to Machine Learning for Engineers. Did you find terms that are new to you? Find out what they mean, and you can create a list of the new terms you encounter.
tähän ladattava PDF, eli komponentti
Ja tähän question scroll tehtäväkomponentti
Tähän vielä kolmanneksi yhdistä oikeat kohdat- tehtävä
You are working as a machine learning engineer in your company. Your boss asks you about machine learning models. Give him/her the right answers
If you wish to read more information about Machine learning, you can click the following link.