Grupo de Ingeniería Estadística Multivariante GIEM
Date established
City
Country
ID
Description
Publication Search Results
Innovation project to validate and select items for assessing transversal competencies in higher education
[EN] The educational improvement innovation project is focused on determining how to evaluate any competence and which are their main related items that should be used for that purpose. Currently, the selection of items is usually performed by groups of experts. However, two main problems related to this type of selection arise in this case: on the one hand, the selection resulting from different groups of experts may be not the same or similar enough, since it is based on the experience and knowledge of each member. On the other hand, the coefficients or weights a priori assigned to each item on any competence invalidate any a posteriori analysis on its statistical significance and the "real" weight on this competence. To mitigate the above drawbacks, this work presents a methodology able to select, from an objective point of view, the items related to a specific competence, from a set of potentially related ones; furthermore the weights associated to the items are determined. This is carried out by applying a multivariate statistical projection method such as Partial Least Squares (PLS), embedded in a cross-validation process. The paper presents how to preprocess the data, analyze it and obtain the items and their weights to be used for the evaluation of a specific competence.
Elaboration process of a map of activities for students in different levels in the competence "Innovation, Creativity and Entrepreneurship"
[EN] The student's skills to innovate, be creative and be entrepreneur are, in most cases, unknown to the teacher when he/she begins his/her teaching activity with a new group of students. In general, the initial level of a student in non-transversal competences can be known by the structure of the previously studied curricula. However, for transversal competences, this starting level is not easy to identify. Furthermore, we found that there are more significant differences between students in the levels of transversal competences than in the levels of non-transversal competences. The diversity of levels in the classroom for the competence led us to propose a map of appropriate activities for each level. The process of elaborating a map of activities for students in different levels in the competence "Innovation, creativity and entrepreneurship" includes the following steps: 1. Identification of activities. The construction of the map requires the identification of learning activities oriented towards the acquisition of the transversal competence. In our case, 11 activities were identified as suitable to work in the "Innovation, Creativity and Entrepreneurship" competence. 2. Definition of the levels. Four categories have been defined to identify the different levels: D. Not achieved; C. Under development; B. Good; A. Excellent. Furthermore, this scale was used for I. Undergraduate students (1st and 2nd year), II. Undergraduate students (3rd and 4th year), and III. Master students. 3. Classification of activities: Data collection. Data from teachers who are experts in the competence should be collected systematically to classify the activities. Thus, each teacher assigned each activity to one or two levels of the scale (from A to D), this for the 3 levels of studies (Level I: 1st and 2nd year, Level II: 3rd and 4th year, Level III: Master). A restriction has been introduced: the level assigned to an activity (A, B, C or D) must be maintained or evolve to lower levels on the scale as we move from level I to level II and level III (due to student progression). 4. Classification of activities: Analysis. The data collected require an analysis process to assign activities at each level. It is necessary to quantify (4 to 1) the qualitative scale (A to D) in order to perform an interquartile analysis. This type of analysis was selected because it is very little affected by extreme values. The median value was taken to obtain the value of the level. A debate was required in case of a high interquartile range. 5. Classification of activities: Map construction. The numerical values of each activity must be transformed to the original scale (A-D). This information was used to build three activity maps, one for each level of studies (I, II and III). 6. Map analysis. The location of the activities on the map allows: on the one hand, a) to identify sets of activities that work at the same level, so it may not be necessary to deploy a whole set of activities in the classroom, since with the completion of one of these activities the students would have already covered the level. And, on the other hand, b) to identify levels not covered by any activity. The process has achieved the established objectives by presenting a map of activities by levels.
Sparse N-way Partial Least Squares by L1-penalization
[EN] N-PLS, as the natural extension of PLS to N-way structures, tries to maximize the covariance between an X and a Y N-way data arrays. It provides a useful framework for fitting prediction models to N-way data. However, N-PLS by itself does not perform variable selection, which indeed can facilitate interpretation in different situations (e.g. the so-called ¿¿omics¿ data). In this work, we propose a method for variable selection within N-PLS by introducing sparsity in the weights matrices WJ and WK by means of L1-penalization. The sparse version of N-PLS is able to provide lower prediction errors by filtering all the noise variables and to further improve interpretability and usability of the N-PLS results. To test Sparse N-PLS performance two different simulated data sets were used, whereas to show its utility in a biological context a real time course metabolomics data set was used.
Prostate Diffusion Weighted-Magnetic Resonance Image analysis using Multivariate Curve Resolution methods
[EN] Multivariate Curve Resolution (MCR) has been applied on prostate Diffusion Weighted-Magnetic Resonance Images (DW-MRI). Different physiological-based modeling approaches of the diffusion process have been submitted to validation by sequentially incorporating prior knowledge on the MCR constraints. Results validate the biexponential diffusion modeling approach and show the capability of the MCR models to find, characterize and locate the behaviors related to the presence of an early prostate tumor.
Statistical Process Control based on Multivariate Image Analysis: A new proposal for monitoring and defect detection
The monitoring, fault detection and visualization of defects are a strategic issue for product quality. This paper presents a novel methodology based on the integration of textural Multivariate image analysis (MIA) and multivariate statistical process control (MSPC) for process monitoring. The proposed approach combines MIA and p-control charts, as well as T2 and RSS images for defect location and visualization. Simulated images of steel plates are used to illustrate the monitoring performance of it. Both approaches are also applied on real clover images.
Prostate functional magnetic resonance image analysis using multivariate curve resolution methods
This paper discusses the potential of Multivariate Curve Resolution (MCR) models to extract physiological dynamics behaviors from Dynamic Contrast Enhanced Magnetic Resonance (DCE-MR) Imaging prostate perfusion studies for cancer diagnosis. A relationship with biomarkers ( hidden parameters for assessing the possible existence of a tumor) from pharmacokinetic models is also studied.
Near infrared hyperspectral imaging for forensic analysis of document forgery
[EN] Hyperspectral images in the near infrared range (HSI-NIR) were evaluated as a nondestructive method to detect fraud in documents. Three different types of typical forgeries were simulated by (a) obliterating text, (b) adding text and (c) approaching the crossing lines problem. The simulated samples were imaged in the range of 928 2524 nm with spectral and spatial resolutions of 6.3 nm and 10 mm, respectively. After data pre-processing, different chemometric techniques were evaluated for each type of forgery. Principal component analysis (PCA) was performed to elucidate the first two types of adulteration, (a) and (b). Moreover, Multivariate Curve Resolution Alternating Least Squares (MCR-ALS) was used in an attempt to improve the results of the type (a) obliteration and type (b) adding text problems. Finally, MCR-ALS and Partial Least Squares Discriminant Analysis (PLS-DA), employed as a variable selection tool, were used to study the type (c) forgeries, i.e. crossing lines problem. Type (a) forgeries (obliterating text) were successfully identified in 43% of the samples using both the chemometric methods (PCA and MCR-ALS). Type (b) forgeries (adding text) were successfully identified in 82% of the samples using both the methods (PCA and MCR-ALS). Finally, type (c) forgeries (crossing lines) were successfully identified in 85% of the samples. The results demonstrate the potential of HSI-NIR associated with chemometric tools to support document forgery identification
Early identification of the student's level in the transversal competence "creativity, innovation and entrepeneurship".
[EN] The initial level of a competence in our students is in many cases unknown for the teacher, when we talk about transversal competences. In addition, in some cases, they assume some skills to the students that can be wrong. The proposal developed in this paper focuses on identifying the student¿s initial level in the competence to adapt the teaching-learning process to the different levels found in our classroom. The proposal has been developed in the competence of creativity, innovation and entrepreneurship.
Proceedings of the VI Chemometrics Workshop for Young Researchers
The Chemometrics Workshop for Young Researchers is a multidisciplinary symposium aimed at gathering young researchers and international experts in the field of multivariate data analysis. The main goal of the Workshop is to allow mainly Ph.D. students to share their knowledge and present the results of their work, encouraging this way the cooperation among different research teams.
PLS: A versatile tool for industrial process improvement and optimization
[EN] Modern industrial processes are characterized by acquiring massive amounts of highly collinear data. In this context, partial least-squares (PLS) regression, if wisely used, can become a strategic tool for process improvement and optimization. In this paper we illustrate the versatility of this technique through several real case studies that basically differ in the structure of the X matrix (process variables) and Y matrix (response parameters). By using the PLS approach, the results show that it is possible to build predictive models (soft sensors) for monitoring the performance of a wastewater treatment plant, to help in the diagnosis of a complex batch polymerization process, to develop in automatic classifier based on image data, or to assist in the empirical model building of a continuous polymerization process.