Python3 Fun Machine Learning (2)

Basic tasks of machine learning

• Classification tasks
• Regression tasks

Classification tasks

Handwriting Input Digital Recognition

Classification tasks:

Two classification tasks

• To determine if the message is spam or not junk e-mail
• Determine the risk or risk of issuing credit cards to customers
• Judging the patient's benign tumor or malignant tumor
• Determine whether a stock is up or down

Multi-category Tasks

• Digital recognition
• Image recognition
• Determine the risk rating of the card issued to the customer.

There are seemingly not categorized tasks that can be converted to categorized tasks, such as using machine learning to play 2048 games and converting the game to move up, down, left, or right in the current state. Another example is the use of machine learning to play Weiqi, automatic driving and so on nursery rhymes can be converted into classified tasks.

2048 games

Some calculation only supports the task of completing the two classification, but the multi-categorized task can be converted to a task of two classification, while some algorithms can naturally accomplish multi-classification tasks.

Multi-label classification, can identify multiple objects in the same screen.

regression task: The result is a sequential numeric value, not a category

• House price
• Market Analysis
• Student Achievement
• Stock price

For the regression task, some algorithms can only solve the regression problem, some algorithms can only solve the classification, and some algorithms can solve the regression problem and solve the classification problem. In some cases, regression tasks can be simplified into categorical tasks. For example, to predict student performance, we may not need to predict the student's specific results and only need to predict whether the student score is a, B, C or D is sufficient, then this is a classification task.

What is machine learning?

After a lot of learning data is fed to machine learning algorithms, the machine learning algorithm will train a model, input samples into the model to get output results, if the model is output a specific category, then we solve is a classification problem, if the model output is a specific value, So what we're dealing with is a regression problem.

Python3 Fun Machine Learning (2)