Kasper Knoblauch Christensen has defended his PhD

Kasper Christensen presented his PhD at April 26th.

Title of the thesis is “Detecting ideas in online communities: Utilizing machine learning and text mining for finding ideas in online communities”. The given topic for the trial lecture was ” What is the future of Big Data in Food, Beverage and Personal Products? – A critical perspective.”

In his thesis, Kasper has explored the possibilities of automatic detection of ideas from online communities.by using text mining and machine learning. He also showed that PLS regression in combination with variable selection can be used to identify the words and phrases that define an idea. He used cases from Lego and beer brewing in his work.

Supervisors:

  • Professor Knut Kvaal, Faculty of Science and Technology, Norwegian University of Life Sciences, Ås, Norway
  • Dr. Argic Einar Risvik, Senior Research Scientist, Department of Sensory, Consumer and Innovation, Nofima, Ås, Norway
  • Professor Tormod Næs, Senior Research Scientist, Nofima, Ås, Norway and Department of Food Science, Quality and Technology, Faculty of Life Sciences, University of Copenhagen, Copenhagen, Denmark
  • Dr. Torulf Mollestad, Principal consultant, Altran, Norway, Oslo

 

Evaluation committee:

  • Dr. Hal Macfie, Visiting Professor, Universities of Reading, Nottingham, United Kingdom
  • Professor Per B. Brockhoff, Department of Applied Mathematics and Computer Science, Technical University of Denmark, Kgs. Lyngby, Denmark
  • Associate Professor, Jorge M. Marchetti, Faculty of Science and Technology, Norwegian University of Life Sciences, Ås, Norway

Visit at Radboud University

Ingrid Måge visited Dr. Jeroen Jansen and his group at Radboud University, Netherlands. Ingrid gave a presentation of Nofima’s work on multi-block data modelling, with special focus on the SO-PLS methodology. Several areas of collaboration were discussed, and we  look forward to working more with Jeroen and his group in the future.

Alessandra Biancolillo awarded PhD

Alessandra Biancolillo defended her PhD thesis at the University of Copenhagen on November 18th. Alessandra has been associated with Nofima’s  strategic program “Multiblock Methods for prediction and interpretation”, developing methodology for analyzing and interpreting several blocks of data sequentially (so-called SO-PLS). The thesis is titled “Method development in the area of multi-block analysis focused on food analysis”. Supervisors have been Tormod Næs (Nofima), Ingrid Måge (Nofima) and Rasmus Bro (University of Copenhagen).

Alessandra did a very good job at the defense. Opponents Thomas Skov (University of Copenhagen), El Mostafa Qannari (Oniris, Nantes) and Lars Nørgaard (Roskilde University) were well prepared and contributed to an interesting and fruitful discussion.

ale_opponents

 

 

Nofima at COST Meeting FoodMC

Kristian Hovde Liland and Ingrid Måge participated last week at the second meeting of the COST action FoodMC. The action is led by INRA, and the meeting took place in INRA’s premises in Versailles. The full title is “Mathematical and Computer Science Methods for Food Science and Industry”. The goal is to Support the food sector in facing futures challenges in production and processing, adopting modelling and optimization methods from Maths and Computer Science.

Ingrid leads the working group called “Modelling food products and food Processes.” This is the largest working group, with over 50 members (and growing). The group is composed of researchers with backgrounds from both mathematical and statistical modeling, two fields that traditionally have limited collaboration. The applications range from mathematical-physical modeling of wafer fracture, through quantification of insects in wheat flour using NIR spectroscopy, to decision support models used to select the appropriate packaging machine. It is an interesting and challenging job to coordinate such a diverse group of scientists, with a correspondingly large potential to expand our horizons and acquire new valuable contacts.

For more info about the COST action, see:

https://www6.inra.fr/foodmc

14th AgroStat symposium

The 2016 AgroStat symposium was held at Nestlé Research Center in Lausanne, Switzerland on March 21-24. It was professionally organized and had a lot of high quality scientific contributions in chemometrics, sensometrics, big data and risk & process. Most of the presentations and all posters were in English, while a few exceptions were French. The first day was reserved for a variety of workshops. There were around 120 participants, mostly French and Swiss, but also from Canada, Great Britain, Scandinavia, Portugal, Italy, USA, Germany and some other countries.

From Nofima Kristian Hovde Liland contributed with a poster describing the R package MatrixCorrelation and the newly developed Similarity of Matrices Index.

Lausanne2

Presentations, posters and short papers are available from:
http://agrostat2016.sfds.asso.fr/e-proceedings/?lang=en

Mini-Arctic 5-6’th November

On November 5’th and 6’th a lunch-to-lunch workshop in chemometrics, statistics and data analysis was held at Nofima, Ås. The workshop was a follow-up on the conference “Arctic Analysis” that was held in Greenland in March 2014. Around twenty people from the Netherlands, Italy, Denmark and Norway  were invited. The participants were a mix of experienced and young researchers, but only the “young and promising” had the opportunity to present their work this time.

The program contained a variety of topics, but the majority were related to data fusion/multiblock methodology. The full titles of the presentations are listed below. The workshop was informal, and the small format allowed good and fruitful discussions. A new mini-workshop is planned next year in the Netherlands, before a full “Arctic Analysis” conference the year after.

On Wednesday evening there was a dinner in Oslo, of course in the “Grønland” district at “Olympen”, where we had traditional Norwegian Christmas food: “ribbe” and “riskrem”.

As usually, Solveig (Adm. Coordinator) had organized the practical parts of the workshop in an excellent way.

Presentations

  • Comparison of structure-revealing data fusion models, Evrim Acar (University of Copenhagen)
  • Procrustes based clustering of multiblock data from projective mapping experiments, Ingunn Berget (Nofima)
  • Variants of Tikhonov Regularization for modelling spectral data, Joakim Skogholt (NMBU)
  • Separating common from distinctive variation, Frans van der Kloet (Leiden University, Netherlands Metabolomics Centre)
  • Analysis of designed experiments with focus on the ANOVA-Target projection approach, Federico Marini ( of Chemistry, University of Rome “La Sapienza”)
  • Classification of plastics containing brominated flame retardant through hyperspectral imaging and chemometrics, Marta Bevilacqua (University of Copenhagen)
  • Multi-set analysis for Multicolour Flow Cytometry: chemometrics for omics analysis of immunology, Jeroen Jansen (Radboud University)
  • A fast opportunistic PLS method for Multiblock data analysis, Kristian Liland (Nofima)
  • SO-NPLS: Sequentially Orthogonalized-Partial Least Squares for multiway arrays, Alessandra Biancolillo (Nofima, University of Copenhagen)
  • How to investigate highly correllated reference data, Åsmund Rinnan (University of Copenhagen(
  • Common an distinct components in data fusion – what does it mean in practice?, Ingrid Måge (Nofima)

Multiblock classification: SO-PLS and LDA

Combining SO-PLS and linear discriminant analysis for multi-block classification

allessandraThe “Combining SO-PLS and linear discriminant analysis for multi-block classification”, written by Alessandra Biancolillo, Ingrid Måge and Tormod Næs was recently published in Chemometrics and Intelligent Laboratory systems. This is the first paper by Ph.D student Alessandra. Congratulations!

Abstract

The aim of the present work is to extend the Sequentially Orthogonalized-Partial Least Squares (SO-PLS) regression method, usually used for continuous output, to situations where classification is the main purpose. For this reason SO-PLS discriminant analysis will be compared with other commonly used techniques such as Partial Least Squares-Discriminant Analysis (PLS-DA) and Multiblock-Partial Least Squares Discriminant Analysis (MB-PLS-DA). In particular we will focus on how multiblock strategies can give better discrimination than by analyzing the individual blocks. We will also show that SO-PLS discriminant analysis yields some valuable interpretation tools that give additional insight into the data. We will introduce some new ways to represent the information, taking into account both interpretation and predictive aspects.

More details

Full reference:

Biancolillo, A., Måge, I., & Næs, T. (2015). Combining SO-PLS and linear discriminant analysis for multi-block classification. Chemometrics and Intelligent Laboratory Systems, 141(0), 58-67.

Available software

The software used in this paper, can be downloaded from software&downloads