Published on 06/06/2019 in Inspire
Because the traditional methods of analysis are no longer enough to analyze today’s huge data volumes. Suffice it to say that by 2025, over 80 billion devices will generate substantial quantities of data via the Internet of Things. You need AI and ML to keep up with your competitors. You streamline your activities, save a considerable amount of money and have a new means of communicating with your customers and employees.
AI and ML are often used wrongly as interchangeable terms. The difference? AI describes a process where computers simulate human intelligence. ML uses algorithms to make models that simulate behavior on the basis of a dataset. So ML is an element of AI. The approach to and application of AI and ML differ from company to company.
With Hyperconnected Infrastructure, you can set up your machine learning without a problem.
How does Proximus put machine learning into practice?
Elena Gil, Lead Analytics Translator: “At Proximus, we have been using a machine learning algorithm since December 2016 to decide whether to send a splicer to resolve a technical issue. The algorithm is fueled by a wide range of data sources with data histories of variable length. The volume of data that the algorithm handles every day is of 500 MB approximately.”
Machine learning has improved the efficiency of our interventions.
Elena Gil, Lead Analytics Translator at Proximus
The algorithm analyzes
“The algorithm is first given the task of defining the conditions to be fulfilled for a cable splicer to intervene. The algorithm connects all the data, finds patterns and calculates the odds that a cable splicer will be necessary to resolve a problem with a customer. These results are sent to our diagnosis tool.”
The diagnosis tool decides
“All our customers have a tag in this tool that indicates the size of these odds. When a customer calls and our tool sees that he or she has a high quota, it automatically decides that we need to send a cable splicer along rather than a member of the intervention service. The intervention data are supplied to the algorithm as well, so that it learns and improves. As a result, we have considerably improved the efficiency of our interventions.”
What is your company’s AI and machine learning experience?
Would you like to know what infrastructure you need to start using machine learning properly? Read our 9 tips and 4 recommendations to get started.