Skip to content Search
Search our website:

Feature engineering for manufacturing sensor data

Feature engineering for manufacturing sensor data

Principal supervisor: Dr Vladislav Ryzhikov

In multiple applications we deal with the data that is a time series data of measurements related to some manufacturing process (e.g., operation of a drilling rig). This data can be used to predict the outcomes of the process (e.g., the quality of the hole produced by the rig). It is typically the case, however, that most of the data generated by the process is irrelevant for the prediction and, if we want to identify features to be used in ML classifiers for the prediction, we need to focus on small segments of the data representing particular temporal events (e.g., the pressure increased, then was flat, and then decreased, which corresponds to the interval of drilling). The focus of this project is to develop a query answering system for feature extraction on process data, that would allow to specify such temporal events and extract their parameters from the data. It is also important to evaluate the efficiency of the system by training ML classifiers based on the features extracted by the system, in various use cases.

Candidate Requirements:

This PhD is aimed at graduates with a strong interest in database technology and machine learning. Prior experience in machine learning is required and in database systems is desirable.

Key References:

[1] Sebastian Brandt, Diego Calvanese, Elem Güzel Kalayci, Roman Kontchakov, Benjamin Mörzinger, Vladislav Ryzhikov, Guohui Xiao, Michael Zakharyaschev: Two-Dimensional Rule Language for Querying Sensor Log Data: A Framework and Use Cases. TIME 2019: 7:1-7:15

Further details about the project may be obtained from:

Principal supervisor: Dr Vladislav Ryzhikov

Further information about PhDs at CSIS is available via this link.