Skip links

Contact

Aigorithmics

Solutions

Anomaly Detection, Predictive Maintenance
(Anomaly Detection, Predictive Maintenance)

We identify atypical values of the analyzed data.

Benefits:

  • Delay reduction in plan implementation.
  • Reduction of time spent on manual checking, studying the cause, and eliminating errors.
  • Thorough analysis in relation to random sampling.

CATEGORY:

Machine learning

INDUSTRY:

heavy industry, thermal treatment

PROBLEM:

Costly method for checking the temperature inside the load – the necessity to drill a fragment of a batch, which is a waste.

SOLUTION:

Artificial intelligence model for monitoring internal temperature of the load.

MAIN EFFECTS OF THE PROJECT:

  1. Cost optimization related to rejected load.
  2. Possibility of additional monitoring and evaluation of the analyzed process (inside the load).

CATEGORY:

Machine learning, Anomalies in text and numerical data

INDUSTRY:

Telecommunication

PROBLEM:

Inconsistency of entries in a database shared between many entities due to changes in systems.

SOLUTION:

Anomaly detection by performing field validation and developing an algorithm for detecting changes in the fields of the JSON object.

MAIN EFFECTS OF THE PROJECT:

  1. Collecting information about particular fields by creating a script for model training.
  2. Retraining an existing model using new (new JSON files) and recalculating statistics.
  3. Anomaly detection in existing fields and finding fields that are significantly similar to the new ones.