How does machine Learning accomplish tasks

The most basic approach to machine learning:

1, is the use of algorithms to parse data

2. Learn from it

3. Make decisions and predictions about real-world events.

Machine learning is "training" with large amounts of data, through various algorithms to learn how to complete tasks from the data.

There are three categories of machine learning:

The first type is unsupervised learning

It refers to automatically finding patterns from information and classifying them into various categories, sometimes called "clustering

problems".

The second category is supervised learning

Supervised learning refers to labeling history and using models to predict outcomes. If we have a fruit and we judge whether it is a

banana or an apple based on the shape and color of the fruit, this is an example of supervised learning.

The last category is reinforcement learning

Reinforcement learning refers to a learning mode that can be used to support people to make decisions and plan. It is a feedback

mechanism that generates rewards for some actions and behaviors of people, and promotes learning through this feedback

mechanism, which is similar to human learning. Therefore, reinforcement learning is one of the important directions of current

research.

There is a difference between machine learning and deep learning, in which a computer's algorithms are able to learn patterns by

finding information from data, just as humans do. Although deep learning is a type of machine learning, deep learning is the use of

deep neural networks to process models more complex, so that the model's understanding of the data is deeper.

Machine learning is generally based on big data. With machine learning, new knowledge can be gained from big data. The general

process is:

Machine learning variable and feature identification

Find key variables or characteristics that affect efficiency,

+ physical process simulation

The mechanism model of state change is constructed, and the precise relationship between input and efficiency in high dimensional

space is obtained.

The agricultural simulation model is a discrete, agent-based relational model.

The energy consumption of the industrial environment requires a fluid model that is continuous, the core is a physical process

simulation, and there needs to be a 3D model of fluid mechanics as an engine

+ predictive modeling

Adjustment and calibration of model output and prediction.