Journal papers

Conference papers

  • Machine Learning for Intrusion Detection in CyberDefense. South Tech Week, Granada (Spain), .
  • , , MBDA in Action. Network Traffic Measurement and Analysis Conference (TMA), Berlin (Germany), . Presentation
  • , , , PARAMO: Enhanced Data Pre-processing in Batch Multivariate Statistical Process Control. Scandinavian Symposium on Chemometrics (SSC16), Oslo (Norway), .
  • , , , X-CAN: Cross-Penalized Component Analysis. Scandinavian Symposium on Chemometrics (SSC16), Oslo (Norway), . Presentation
  • , All sparse PCA models are wrong, but some are useful. Scandinavian Symposium on Chemometrics (SSC16), Oslo (Norway), . Presentation
  • Multivariate Statistical Control in the Big Data Context. II Arctic Analysis, (Iceland), .
  • , , , Parameter stability and its effects on bilinear modelling of batch processes. Mini Arctic Workshop, Valencia (Spain), .
  • , , Cross validation in Sparse PLS. Mini Arctic Workshop, Valencia (Spain), .
  • , , Orthogonal Procrustes Problem to generate multivariate datasets by simultaneously controlling the covariances structure and the rows distribution. Mini Arctic Workshop, Valencia (Spain), .
  • , GPCA for improved multivariate analysis interpretation in lipidomics. Wenvomics, Barcelona (Spain), .
  • , , , Network-wide intrusion detection supported by multivariate analysis and interactive visualization. VizSec, Phoenix (USA), .
  • , , A UNIVARIATE APPROACH FOR DIAGNOSIS IN PCA-MSPC. Scandinavian Symposium on Chemometrics (SSC15), Naantali (Finland), .
  • , , GROUP-WISE PRINCIPAL COMPONENT ANALYSIS. Scandinavian Symposium on Chemometrics (SSC15), Naantali (Finland), . Presentation
  • , , Traffic Monitoring and Diagnosis with Multivariate Statistical Network Monitoring: A Case Study. IEEE Security & Privacy International Workshop on Traffic Measurements for Cybersecurity (WTMC 2017), San Jose, California (USA), . Presentation
  • , , Exploratory Analysis on Big Data using the MEDA ToolboxPresent and Future. Mini Arctic Workshop, Groningen (Netherlands), . Presentation
  • , , , Hierarchical PCA-Based Multivariate Statistical Network Monitoring for Anomaly Detection. 8th IEEE International Workshop on Information Forensics and Security (WIFS), Abu Dhabi (UAE), .
  • , , , , On the Feasibility of Distinguishing Between Process Disturbances and Intrusions in Process Control Systems using Multivariate Statistical Process Control. The 3rd International Workshop on Reliability and Security Aspects for Critical Infrastructure, Tolouse (France), .
  • , Visual Steering in Multivariate Exploratory Data Analysis. 16th Chemometrics in Analytical Chemistry, Barcelona (Spain), .
  • , , Fault Diagnosis : Contribution plots Vs oMEDA. 16th Chemometrics in Analytical Chemistry, Barcelona (Spain), .
  • , , , , DRNS: Dynamical Relay Node Placement Solution. PAAMS, Sevilla (Spain), .
  • Multivariate Big Data Analysis and its Application on Internet. Invited Keynote: 16th Chemometrics in Analytical Chemistry, Barcelona (Spain), . Presentation
  • , , , Exploratory Analysis in Big Data based on PCA and PLS. Invited Keynote: 14th Scandinavian Symposium on Chemometrics, Sardinia (Italy), . Presentation
  • , , , Tackling the Big Data 4 Vs for Anomaly Detection. INFOCOM'2014 Workshop on Security and Privacy in Big Data, Toronto (Canada), .
  • , , A Multiagent Self-healing System against Security Incidents in MANETs. Workshop on Active Security through Multi-Agent Systems (WASMAS), Salamanca (Spain), .
  • , , , Tampered Data Recovery in WSNs through Dynamic PCA and Variable Routing Strategies. ICCNS - Journal of Communications, Vol. 8, November, (), , pp:738-750.
  • , , A Security Response Approach Based on the Deployment of Mobile Agents: A Practical Vision. 11th International Conference on Practical Applications of Agents and Multi-Agent Systems (PAAMS 2013), Salamanca (Spain), .
  • , , A Security Response Approach based on the Deployment of Mobile Agents. 11th International Conference on Practical Applications of Agents and Multi-Agent Systems (PAAMS 2013), Salamanca (Spain), .
  • , , , , Pair-wise Similarity Criteria for Flows Identification in P2P/non-P2P Traffic Classification. The Third International Conference on Advances in P2P Systems, Lisbon (Portugal), .
  • , , Exploratory data analysis: a latent subspace approach. 5th International Chemometrics Research Meeting, Berg en Dal (Netherlands), .
  • , , Issues on Batch Multi-Variate Statistical Process Control. Eastern Analytical Symposium and Exposition, Somerset (USA), .
  • , , Crossvalidation in Principal Component Analysis: searching for the best approach. 12th Conference on Chemometrics in Analytical Chemistry, Anvers (Belgium), .
  • , , , Application of missing data methods for exploratory data analysis in medical research. VII Colloquium Mediterraneum, Granada (Spain), .
  • , Principal Component Analysis for very large data sets. VII Colloquium Mediterraneum, Granada (Spain), .
  • , , Data Understanding with Principal Component Analysis. 11th Scandinavian Symposium on Chemometrics, Loen/Stryn (Norway), .
  • , , Covariance Maps for Batch Process Modelling. 11th Scandinavian Symposium on Chemometrics, Loen/Stryn (Norway), .
  • , , Multi-phase analysis framework: pattern recognition for batch process modelling. IFPAC, Baltimore (USA), . Presentation
  • , , Run-to-run optimization of fed-batch processes with unfold-PLS. IFPAC, Baltimore (USA), . Presentation
  • , , On-line monitoring of batch processes: Does the modelling structure matter?. 11th Conference on Chemometrics in Analytical Chemistry, Montpellier (France), . Presentation | Abstract
  • , , Leave-n-Samples-Out Cross-validation in PCA for Missing Data Recovery and Robustness in front of Measurement Noise. 11thconference on Chemometrics in Analytical Chemistry, Montpellier (France), . Poster | Abstract
  • New Methods Based on the Projection to Latent Structures for Monitoring, Prediction and Optimization of Batch Processes. Invited Presentation. 9th Belgian Chemometrics Symposium, Gembloux (Belgium), . Presentation
  • , , New advances in the on-line monitoring of batch processes. IFPAC, Baltimore (USA), . Abstract
  • , , A new algorithm for selecting the unfolding method and the number of sub-models in batch process modelling with PCA. 10th Scandinavian Symposium on Chemometrics, Laapperanta (Finland), . Poster | Abstract
  • , , New Cross-Validation Methods in Principal Component Analysis. 10th Scandinavian Symposium on Chemometrics, Laapperanta (Finland), . Poster | Abstract
  • , , A new look at the dynamic covariance structure of various approaches for batch process modelling. 10th Scandinavian Symposium on Chemometrics, Laapperanta (Finland), . Poster | Abstract
  • , , Multi-Phase Analysis Framework for Handling Batch Process Data. 10th Scandinavian Symposium on Chemometrics, Laapperanta (Finland), . Presentation | Abstract
  • , , , Implementing a Cooperative Framework among Bio-inspired Robots based on Phonotaxis. Multi Agent Robotic Systems, International Conference on Informatics in Control, Automation and Robotics, Barcelona (Spain), , pp:31-38.
  • , , Control Adaptativo para el Seguimiento de Robots. Workshop en Agentes Físicos (Congreso Español De Informática), Granada (Spain), , pp:101-108.
  • , , , , Simulation of mobile robot applications with VirtualRobot. International Industrial Simulation Conference, Berlin (Germany), , pp:68-72.


Download José Camacho Thesis

The present Thesis is devoted to the study and application of new methods based on the projection to latent structures to batch processes. The potential economic profit and safety and ecological benefit in the improvement of the modelling, monitoring and control of these processes is widely accepted in both the academic and industrial fields.

The document has been structured in three blocks. Firstly, the related literature is revised and summarized in an introductory part. The principal projection to latent structures methods are presented using a comprehensive common nomenclature. Also, the main contributions in the literature for the modelling, monitoring and control of batch processes are discussed.

In the second block of the document, the general features of batch process data and the basis of the methods based on the projection to latent structures are studied. The study, as the rest of the Thesis, is limited to bilinear modelling approaches. This allows to observe the limitations of these modelling methods when applied to batch processes, some of which are stated here for the first time. In this block, a first theoretical analysis is carried out to investigate the dynamic covariance structures of different modelling approaches. The conclusion of this analysis is that the multi-phase modelling structure has a nice feature which makes it specially attractive for batch process modelling: the flexibility to adjust to the process nature. Multi-phase and single-phase processes, with short or long term dynamics -or even combination of both- can be effectively modelled with the multi-phase structure. The second step in this block of the document is to investigate one of the most widely used strategies to set the parameters of the projection to latent structures methods: cross-validation. New algorithms for cross-validation are proposed and extended to the batch process data case. This step is of vital importance for the development of an automatic modelling algorithm for batch data. This algorithm is presented in the last part of this second block of the Thesis. The algorithm, named Multi-phase algorithm, has been designed together with a set of tools to analyze and handle multi-phase models: the Multi-phase Framework. This framework is a modelling approach which combines the automatic recognition capabilities of the algorithms from the field of Computer Science with the well-founded data analysis from the field of Statistics.

After the study of the modelling of batch processes, several applications where this modelling is necessary are investigated in the last block of the document. The problems addressed are: the off-line and on-line monitoring, the on-line end-quality prediction, the on-line estimation of batch trajectories and the process optimization. All these problems share the common feature that batch processes are difficult to understand and to model. Nonetheless, each of the problems studied presents its particularities, which need to be addressed independently. The contributions in each of the applications go beyond the use of the new modelling approach. In both the off-line and on-line monitoring of batch processes, several improvements in the development of the monitoring system are suggested. In the on-line end-quality prediction and estimation of trajectories, a broad comparison among several modelling approaches is supplied. Finally, a new optimization algorithm, based on the combination of self-tuning extremum seeking and projection to latent structures methods, is proposed.