5 minutes to learn about machine learning

What machine learning is?

Humans have been trying to make machines smart since they were invented. In the 1950s, the development of artificial intelligence experienced a "reasoning period", by giving machine logic reasoning ability to make machines intelligent, but because of its lack of knowledge, far from achieving real intelligence, and human knowledge is huge, one by one transmission to the machine is not realistic, if the machine can learn from the problem is not solved? So machine learning came into being. Machine learning specializes in how computers simulate or implement human learning behaviors to acquire new knowledge or skills, reconstruct existing knowledge systems, and continuously optimize their performance. As a multidisciplinary interdisciplinary subject, machine learning is not only the core of artificial intelligence, but also the fundamental way to make computers intelligent.

Classification of machine learning

Machine learning research and construction is a special algorithm that is not a particular one, allowing the computer to learn from the data itself to make predictions, so machine learning is not a specific algorithm, but a general term for many algorithms, we often hear that deep learning is one of them, other methods are decision trees, clustering, Bayes and so on. Over the decades, studies have published a wide variety of methods of machine learning that can be categorized according to concerns: Based on learning strategies: Machine learning that can be classified as simulating the human brain and machine learning that uses mathematical methods directly. Based on learning methods: Can be classified as generalized learning, deductive learning, analogy learning and analytical learning. Based on learning methods: Can be classified as supervised learning, unsergency learning, and intensive learning. Based on data form: Can be classified as structured learning and unstructured learning. Based on learning objectives: Can be classified as concept learning, rule learning, functional learning, category learning, Bayes e-learning.

Several typical machine learning algorithms

Linear regression algorithm: If there is a "linear relationship" between 2 or more variables, then the historical data can be used to find out the "set" between variables and build an effective model to predict future variable results. Linear regression, which was originally a concept in statistics, is often used in machine learning.
Decision Tree Algorithm: It is a simple machine learning algorithm, which is usually used to solve classification problems, adopt tree structure, and use layer reasoning to achieve final classification.
K proximity algorithm: The core idea is that if a sample in the feature space k most of the most adjacent samples belong to a category, then the sample also belongs to this category, and has the characteristics of the sample in this category.
K mean clustering algorithm: First randomly select K objects as the initial cluster center, and then calculate the distance between each object and each seed cluster center, assign each object to the nearest cluster center, cluster center and the object assigned to them represent a cluster, once all objects are assigned, each cluster cluster center will be recalculated according to the existing objects in the cluster, the process will be repeated until a termination condition is met.

The basic idea of machine learning

No matter what algorithm is used, machine learning ideas can hardly jump three steps: Abstract real-life problems into mathematical models and be able to figure out the role of different parameters in the model. Use mathematical methods to solve the mathematical model, so as to solve real-life problems.
Can evaluating this mathematical model really solve real-life problems? How's the solution? When the idea is clear, you can carry out practical operations, machine learning in the practical level of the process can be summed up simply to collect data
→deste→ select a model →training→ evaluation→ parameter adjustment→ prediction (starting to use). Of course, in the course of practical operation, we will find that not all problems can be converted into mathematical problems, in other words, for those who have no way to convert to mathematical problems of the real problem, AI often has no way to solve, and the most difficult part of machine learning as a whole is actually how to turn real problems into mathematical problems. Machine learning is one of the hottest technologies in the industry, all kinds of artificial intelligence applications have the components of machine learning efforts, we enjoy the convenience of IT technology at the same time, you can learn more about machine learning related knowledge and concepts.

What is Precision Marketing?
Big Data: Basic Concept
Machine Learning- Algorithm Summary