MAX IV Staff R&D. Machine learning applied to natural language processing. Possibilities for MAX IV.

by Mr Isak Lindhe

Wednesday 14 Mar 2018, 13:00 → 14:00 Europe/Stockholm

MAX III meeting room (MAX IV Laboratory)

Title

MAX IV Staff R&D. Machine learning applied to natural language processing. Possibilities for MAX IV.

Agenda

• Introduction and Motivation
• Isak Lindhe. Classification of User Reviews and Identification of Problems. (MAX IV & LTH Project: Visualizing MAX IV User feedback and sentiments)
• Max Söderman. Defining new project using machine learning tools applied to MAX IV users publications.
• Exploring further interest for MAX IV. Collect feedback from the audience.

Classification of User Reviews and Identification of Problems. Isak Lindhe, LTH.

MAX IV aims to provide efficient user support to contribute to high quality scientific output. One of the many essential aspects is to collect and act on the feedback provided by the users after their experimental sessions. This aims at contribute to optimize the operation of the facility performance, the users research outcome as well as the user experience.

The analysis of MAX-lab user feedback, within this project, was based on natural language processing techniques to categorize text, detect topics, and analyze sentiment. The work required to determine the most relevant algorithms for these domains, experiment and evaluate them, and possibly adapt them. The development of the application was built on an existing corpus that the students use to create machine learning models used by the algorithms.

Here it will be presented the attempt to use natural language processing to extract useful information from user feedback and display it in a web application.

The work was done as part of the LTH course Project in computer science.