New ways to learn (Machine learning)
Before starting with the basic concepts of machine learning, it is important to enter into context with a simple example of this beautiful world of machine learning.
Let’s answer a simple question, what is a dog may seem silly but it is difficult to give an exact description of what it is at the time with this I mean when we see a dog for the first time, it is very similar as when we Question that is a tree Everyone knows what a tree is but in the first instance the definition and interpretation of what a tree is not present if not later than when under experience and recognition we identify what a tree is, We all know what a tree is but finding the right definition is not.
A clear example of learning and understanding to identify what a tree is well illustrated in the following image, where we can see what a tree is but other elements are similar but at the time we do not know for sure if it is also a tree or not and To do this, we must ask someone else who has that experience and knowledge necessary to be able to instruct whether or not it is a tree and based on this, we can correctly identify what a tree is and what is not.
Now if we go to the exact definition of what machine learning is and how it works in the world, it consists of identifying and learning data patterns to be able to make predictions is simply a set of algorithms that perform this act of learning, for this you have to train a model, in this case the model refers to that person who initially tries to learn to identify what it is and that it is not a tree at the software level from there you can leave only the software to make your own decisions.
It is important to remember that there are currently other concepts that sound very similar but are not the same, for example:
Big Data, consists of the daily study of your data on the internet of how you move on the network, what you do and how you do it, this is a very general definition since the exact definition goes a little further and consists of The manipulation of large volumes of data that cannot be used in a conventional manner and the objective of this is to be able to make decisions in the short, medium and long term on issues related to the sale of products or services.
Data mining, Data mining is precisely responsible for discovering patterns in large amounts of data that is why it is mainly used in statistics and computer science many believe that Data Mining is responsible for collecting, extracting and processing data but the reality is that responsible for debugging existing data and finding relevant and interesting patterns.
Artificial intelligence or AI, It is the simulation of human intelligence processes by the computer system that include learning and decision making, artificial intelligence is based on understanding human capabilities either for specific tasks or for larger tasks and trying to improve Each of these processes to expedite these tasks, such as mathematical analysis, interpretation of human language and even make their own decisions based on the information they are interpreting.
Deep learning, is an extension of the machine learning area but it is a much deeper and more complex area where the study of the data is deepened to a more complex level, this is presented because every day the algorithms responsible for the analysis and interpretation of the data grows and grows exponentially resulting in this new area of machine learning known as Deep learning.
And to conclude a bit on the subject and deepen the topic Machine learning let’s make some differences clear.
Artificial intelligence or AI its main objective is to ensure that machines or in particular computers manage to develop some of the tasks that humans usually develop a fairly futuristic theme that is very present today, on the other hand Machine learning is an area of the computer science much more applied where what is done is to develop algorithms that are capable of taking some data and from these numerical or digital data stored in the computer to extract some characteristics or parameters of interest and to be able to generate some data at the output that allow me to interpret the information I am entering into that system. Deep learning is specifically a smaller area of machine learning where we use algorithms very similar to human neural networks.
Why is there a need to use the technologies that are developed in the Machine learning area and what is your main objective in this new digital era?
In general, it could be said that the world is undergoing an incredible transformation in the field of the use of data found today on the internet, for this reason an efficient way is created to obtain this data that travels on the internet and then give it a commercial, social, educational and even in the area of health, so that we get an idea in the world every day about 50 Million Tweets are made and around 900,000 posts are published daily, this large amount Information is incredibly valuable to one who for some people does not believe that it is relevant but the reality is that if it has an enormous value that tells us very accurately for where our society is heading in all areas that the human being develops.
The main objective beyond the economic end that this generates for all the companies that make use of this source of information is to correctly achieve behavioral patterns that help us improve every day as human beings.
How machine learning works and its algorithms?
Every process that requires an analysis and a subsequent conclusion, it is necessary to follow different stages that will lead us to a good result.
The first thing we have to do is take all the data that we have available so far to be able to analyze this information. It is very important to keep in mind that these data to analyze contain a large volume of information in this way we ensure that we cover all the possibilities of We want to analyze and conclude.
Once we have all the information we need to analyze we must prepare all this information at this stage it is known as data preprocessing where we clean the data of little relevant information for the analysis that is planned for said study, once this process is ready clean the data we make sure to define what kind of algorithm to use for the analysis of this data.
The next stage will be to determine which algorithm is the most suitable to analyze all the information, now we will see which are the most commonly used algorithms and we have three types the first is Linear, Tree based and finally Neural networks of each of these types are It gives off other sub categories.
Sources
Deep learning is an extension of the machine learning area but it is a much deeper and more complex area where the study of the data is deepened to a more complex level, this is presented because every day the algorithms responsible for the analysis and interpretation of the data grows and grows exponentially resulting in this new area of machine learning known as Deep learning.