Distributed Artificial Intelligence Models for Knowledge Discovery in Bioinformatics

Juan M. Corchado, Isabelle Bichindaritz, Juan F. De Paz
2015 BioMed Research International  
The increasing volume of existing information on biological processes and the use of large databases have significantly increased the accessibility of datasets to the scientific community. This has enabled performing an analysis to facilitate the extraction of relevant information or modeling and optimizing tasks in different processes. Parallel to the increasing volumes of information is the emergence of new or adapted distributed computing models such as grid computing and cloud computing.
more » ... ether with new techniques of artificial intelligence, or more specifically knowledge discovery, these management systems are making it possible to perform a more efficient analysis of the information and are enabling the creation of adaptive systems with learning ability. In the area of distributed artificial intelligence models for knowledge discovery in bioinformatics, ten interesting proposals are presented. These models analyze different biological aspects and simulate the process or user behavior in the health care system. The main characteristics in these proposals are the use of artificial intelligence techniques to analyze the information and extract knowledge. "Bladder Carcinoma Data with Clinical Risk Factors and Molecular Markers: A Cluster Analysis" provides interesting research about bladder cancer. The paper shows the hypothesis that the use of clinical and histopathological data with information about marker is useful to manage treatments of nonmuscle invasive bladder cancers (NMIBC). The authors apply data mining techniques such as hierarchical clustering to create molecular cluster and risk groups. In their experiments, the authors analyze 45 patients with a new diagnosis of NMIBC. They create four groups of patients and categorize the patients according to clinical characters and biological behavior. The authors of "A Linear-RBF Multikernel SVM to Classify Big Text Corpora" use data mining techniques based on classifiers to big text corpora. In particular, they implement a variant of support vector machine (SVM) to reduce the computational cost. The authors show a multikernel SVM with automatic parameterization to improve the results of SVM parameterized under a brute force search. The proposal is composed of a workflow with algorithms to process documents, reduce the dimensionality of the data, and to apply/ provide clustering, training, and prediction. The proposal is analyzed according to the classification results and building time in the dataset TREC Genomics 2005 corpus. In "Analysis of Environmental Stress Factors Using an Artificial Growth System and Plant Fitness Optimization, " the authors analyze how some environment conditions can accelerate the evolution process; in addition to modifying the environment conditions, it is possible to select genomic variants. The authors analyze several factors for Pleurotus ostreatus to improve quality and production. In order to carry out the analysis of the factors, the authors include several IoT sensors to retrieve the information from the sensors. The information is then processed with data mining techniques
doi:10.1155/2015/846785 pmid:25883977 pmcid:PMC4389817 fatcat:ext77dryo5azlovalv356v3rvu