Algorithms to be used in daily work

Machine Learning are not a magical tricks anymore

The Artificial Intelligence we hear so much about is actually the use of algorithms. The phrase "Machine Learning" (ML) and the statement that computers learn from the data they receive are common. In fact, computers do not learn anything - they are soulless machines that only execute commands. Computers provide us with the results of their work in two ways; we tell the computers an example of what to look for in the data (supervised learning) or we tell them to run an algorithm that has predefined results to deliver ("unsupervised learning").

Read more about types of algorithms and their appliance.

Thus, a finite number of algorithms can give us an infinite number of combinations of their use – it is amazing how short is the list of functions and algorithms that create quite complex software that performs various tasks. It's like the language we use; the number of words is limited, but the number of literary works will never reach the end.

What are algorithms?

Algorithm a finite sequence of clearly defined actions necessary to perform a certain type of task. A recipe is often cited as an example of an algorithm used in everyday life.

ML uses algorithms that come from the 60s, 50s and even ancient Greece. Their wide use was made possible by fast computers. It is fast computing machines that have made a real revolution, not the algorithms themselves.

The use of algorithms was not overly difficult, but so far it required some skills, like coding ability. However, that has changed. Since some time we have access to software where we have algorithms on the right button of a mouse. What's more, there is free software like Orange Data Mining - a great environment in which we can test certain hypotheses, look for hidden information in data, get recommendations on the basis of which we will make further decisions. Here are several examples of several "Machine Learning" algorithms that you can use in your daily work.

Linear regression

Linear regression and especially the model that was built on that algorithm - Auto Regressive Integrated Moving Average (ARIMA). This is a statistical analysis model that uses time series data to either better understand the data set or to predict future trends.

What it can be used for:

  • Stock price prediction
  • Supply chain forecasts
  • Procurement and production planning

FP Growth

FP-growth is an association algorithm for frequent pattern, associations, or causal structures within the set of data. For example, grocery store transaction data might have a frequent pattern that people usually buy milk and beer together.

What it can be used for:

  • Basket data analysis – To analyze the association of purchased items in a single basket or single purchase.
  • Cross marketing/Selling
  • Recommendation systems – e-commerce site or CRM features
  • Medical Treatments – Each patient is represented as a transaction containing the set of diseases, and which diseases are likely to occur simultaneously or sequentially and can therefore be predicted.

Hierarchical clustering

Hierarchical clustering, also known as hierarchical cluster analysis, is an algorithm that groups similar objects into groups called clusters. The algorithm is based on a matrix of distances between objects. The distances are calculated according to another algorithm, such as the Euclidean algorithm [many others possible]

What it can be used for:

  •  Review analysis - Opinions often have common elements of content
  • Marketing and Sales - Personalization and targeting in marketing is big business.
  • Classifying network traffic - Group together characteristics of the traffic sources

If you want to know more about what else you can do using algorithms, visit the Orange website. There you will find more examples and tutorials.

So long!