What’s on our mind?

Exsilio Solutions is proud to introduce our new blog! Not only is this forum completely integrated with all our social networks, but it is also 100% responsive. Now, you can take Exsilio with you on your phone, tablet, and desktop - redefine what you thought possible!

Introduction to Machine Learning

Machine learning is an analysis of data to use it to drive answers. Data is derived by detecting statistical patterns and trends.

MACHINE LEARNING

What is Machine Learning?

“The Field of study that gives computers the ability to learn without being explicitly programmed.”

Arthur Samuel 

We learn through our past experiences. But what about computers? Can computers get to learn from the experience too?

The answer is YES!! That exactly is Machine Learning. Computers experiences are called as data.

Machine learning is an analysis of data to use it to drive answers. Data is derived by detecting statistical patterns and trends.

How is Machine Learning useful in Business Intelligence?

Machine Learning is all about turning information to knowledge. We are generating a massive amount of data nowadays and from that data coming up with Static BI report is not the solution. Insides of that data are too deep and meaningful which is giving us a lot of information. From that Machine Learning comes to the picture. Machine Learning techniques help us to automatically find valuable patterns within complex data.

Machine learning process:

 

 

 

 

 

 

Machine Learning is the intersection of Artificial Intelligence and Data mining techniques.

It seeks to use algorithms to derive meaning from DATA.

Types of algorithms

Two of the most widely adopted machine learning methods are supervised learning and unsupervised learning – but there are also other methods of machine learning. Here's an overview of the most popular types.

Supervised

Target values are known

Supervised learning is the easiest to understand and simplest to implement. It is the way to train a machine with labeled data. Which means we already have the output of the given input dataset. For Example, let's provide a computer series of features which are columns of data like Age, Height and that sort of things and then we try to predict the weight of a person using Machine Learning algorithm. After training an algorithm using a Training set it is ready to use the past experiences and begins to understand that the older someone or the taller someone is more likely to have the greater weight. So, supervised learning is about training an algorithm just like teaching a human child.

Types of Supervised learning:

  • Classification: Identify the most likely outcome from the multiclass list or assigning a category to each item. For example, Hotels’ menu is a classification of items. Get a new menu and try to predict which kind of restaurant it is (Indian, Italian, Chinese, Mexican). This works all down to image recognition also. Classification is all about asking a question Which Category
  • Regression: Take the past and try to fit it into trend line to predict the future value (Stock value, currency value) It deals with How much /How many you are working with.

Below are the key points for supervised learning:

  • Labels provided
  • Learn from known past example to predict the future
  • Cross-checking is possible with model’s predictions against the true outputs

Unsupervised

Target values unknown

In Unsupervised learning, only input data is provided. There are no labels available to learn from and cross-check the output is true or not. Unlike supervised learning, unsupervised learning is difficult as no training set is available. The machine will find patterns and relationships from the data and looking to group things together. In some cases, simply examine and see how they relate to each other and how strongly they relate to each other.

  • Clustering: (Which group) Here we are looking to do the partitioning items into Homogeneous groups, such as grouping customers by their purchases or taking social media posts and putting them together (making groups of Twitter, Instagram, Facebook posts).

Below are the key points for unsupervised learning:

  • No labels provided
  • Learning the structure of data

Any thoughts?



Loading more content...