Big Data 4/6
Welcome to the fourth entry about the big data, I hope to subscribe, interact and give you Resteem. Thanks @javf1016 :)
First Entry: https://steemit.com/business/@javf1016/big-data-1-10
Second Entry: https://steemit.com/education/@javf1016/big-data-2-10
Third Entry: https://steemit.com/education/@javf1016/big-data-3-10
III. Machine Learning in Computer Security
Machine Learning, automatic learning or machine learning, is one of the components of artificial intelligence focused on applying mathematical models so computers, machines can learn. To extract the environment, the real nature in which we find ourselves, interpret it and incorporate it into a mathematical model, results in a machine language, which can be read and interpreted by artificial elements, turning them into collaborators with the capacity for self-learning.
Learning is totally linked to generalization, to the ability to react in different contexts; We must take into account that in this context we will need (3) three elements mainly, which are:
Agent
Environment
Function
The agent is an entity that is able to perceive what happens in the environment for which it is developed and can interact with it. Fig. 6. The function is responsible for perceiving the normal behavior and detect any anomaly, unusual process within the system in real time.
Fig. 6 Agent interacts with the environment to acquire knowledge
An agent acquires knowledge, learns if he gains the ability to change his behavior in real time, improving the function he has dedicated, bringing it to its optimal state. The acquired learning or knowledge is classified into (3) three major types:
Unsupervised Learning: The agent does not need external information to which it is handled.
Supervised Learning: There is a supervisor who enters information to manage.
Reinforcement learning: There is a supervisor who introduces information to the agent and verifies if what is done is good or bad, but does not change its function.
The large daily volume of alerts generated in a system causes human beings to lose efficiency in the analysis, something that a machine can use in its favor, making its predictive capacity increase. For this process we will introduce a new term which is "Data Mining", process in which we extract large amounts of data in order to generate patterns and thus identify and establish relationships.
A pattern recognition system contains:
Sensors
Interpretation
Prediction
Learning
In the following image we observe how different systems are suitable to the exposed patterns. Fig. 7
Fig.7 Learning process
Knowledge is shaped in the function of multiple forms since they are mathematical functions, it can be mapped in the form of tree, network, probabilistic model, geometric model, etc. The application of the learned model is in a process called classification, is within what we call interpretation, what the agent performs is to query the model or function by an input vector, the model will perform a classification of what comes through Vector and associate it with one of the learning it has and finally the agent will be able to make a prediction and act according to what it establishes.
Glossary
Machine learning, automatic learning.
It's very interesting post!
Exchellent job dear @javf1016 :)
Thanks @djsonic