Advanced process analytics, TrendMiner

Descriptive analytics

To be able to define the What?, that is to say, what happened in the process? Which state of the process is representative of the problem, select interesting or abnormal patterns, relevant events or certain events?

This is thanks to the filters by time, criteria available in TrendMiner as well as the ability to create formulas and import data.

Discovery Analytics

In order to define the When?, How often?, that is, if the event has happened before and how often.

Define where it happened, under what context, how many times?, facilitate the generation of a hypothesis and be able to test it. This thanks to the algorithm of TrendMiner of search by similarity, in a graphical way TrendMiner searches if the same behavior pattern has been presented previously.

Also, through the search by value, which allows specifying criteria for an established period as well as search by area of operation that allows defining under what conditions the process has operated.

Diagnostic analytics

In order to define the Why?, used to find the root causes or generate a hypothesis.

To be able to define Why did it happen?, to compare against the normal operation of the process, to be able to define what factors influenced this situation.

This is thanks to the ability of TrendMiner to compare multiple charts stacked in layers, for example to compare "good" periods against "bad" periods. As well as the Algorithm of Recommendations and Influence Factors that generates suggestions of root cause variables of the anomaly based on a linear relationship.

Predictive analytics

To be able to define What’s next?, To be able to predict the behavior of the process, in which way an anomalous situation can evolve.

This is thanks to the Fingerprint tool that allows creating a "Fingerprint" of the optimal conditions of the process once defined, this fingerprint can be compared against the process data in real time in order to detect deviations in the process early; with the ability to generate notifications and alerts when these deviations are detected by TrendMiner through the Predictive Mode option.