Predictive Analytics in the Supply Chain
We are currently in a new interconnected era where the economy begins to be based on a resource that is not only renewable, but also self-generated.
Internet of Things (connected systems) allows flows of large volumes of data at high speed.
TODAY, it is not only possible to capture the machines’ behaviour but we also do it in real time.
And according to Peter Drucker, “What can be measured can be managed”.
BUT HOW? HOW DO WE BENEFIT?
Predict Analytics, like BI, uses mathematical and statistical algorithms to analyze historical data. The difference is that Predictive analytics looks at the prediction.
It is no longer necessary to know the rules (or conditions) that govern the system. Intelligence is no longer imposed by humans (who define the rules), because Predictive Analytics learns automatically according to the history. In this case, the intelligence is given according to the experience that has been given to the model automaticly.
Predictive analytics (based on representing the information on a numerical way, which) allows finding a function that contains the rules (patterns) that will allow the prediction.
THE PROCESS IS CARRIED OUT IN TWO PHASES: ANALYSIS OR TRAINING, AND IMPLEMENTATION. THEN, YOU HAVE TO KEEP IT.
During the construction of the model, the analysis (or training) is carreid out, where the importance of each variable and the relationships between them are sought, locating patterns that will later allow us to obtain information about future events.
When the model is already trained, we implement the model in the information flow, analyzing all data in real time.
However, the current world changes fast, so it is necessary to maintain our models to improve them as new information arrives.
But all this technology is useless if you do not have questions. In Predictive Analytics the question is the center.