Improvement in the early detection of cancer breast.
a study at the Polytechnic University of Madrid achieved a significant improvement in the classification of breast cancer data
Spain, 2011-December Researchers of the Automation Group in signal and communications from the Universidad Politécnica de Madrid (CSAG/UPM) implemented a new learning method for artificial neural networks based on the synaptic metaplasticidad of biological neurons, which has allowed to classify patterns database Wisconsin (WBCD) breast cancer, international reference in mammography, with an accuracy of the 99.63%.
Cancer is one of the main causes of mortality worldwide and research into the diagnosis and treatment has become a topic of vital importance to the scientific community.
Prevention remains a challenge, and the best way to increase the survival of the patients is through early detection. If cancer cells are detected before its spread to other organs, the survival rate is higher than 97%.
For this reason, the use and development of automated classifiers to give support to medical diagnosis has risen markedly in recent times. These classification systems attempt to minimize errors produced by specialists, increase the number of diagnostics that can perform in a given time, and its success rate. Most of these systems are based on techniques of artificial intelligence combined with signal processing, mainly: artificial neural networks, wavelet analysis, image analysis using models Bayesian, machines vector support, fuzzy logic and fractal models among other powerful mathematical techniques.
is specifically an artificial neural network (AMMLP), trained with a new method (Metaplasticidad Artificial) proposed by Professor Diego Andina and applied to data from cancer by the researcher Alexis Marcano-Cedeño, both belonging to the Group on automation in signal and communications of the Polytechnic University of Madrid (CSAG/UPM)which has achieved the best results so far.
Metaplasticidad
The concept of biological Metaplasticidad was defined in 1996 by W.C. Abraham The prefix goal ” comes from the Greek beyond ” or above ” while the word plasticity ” relates to the ability to have neurons change the value of the strength of synaptic joints. Abraham defined the metaplasticidad as the induction of synaptic changes in function of the previous synaptic activity, i.e. the metaplasticidad depends on much of the history of activation of the synapses, and hypothesized that the metaplasticidad plays an important role in stability (homeostasis)learning efficiency and mechanism of biological memory.
this database is one of the best known and used for testing algorithms of classification of patterns of breast cancer. The WBCD consists of 699 samples. Each record in the database has nine attributes. Integer from 1 to 10 values are assigned to the assessments, 1 being the closest benign and 10 closest to evil. Each sample is also associated with a tag of class, which can be benign ” or evil ”. This data set contains 16 entries with attribute values missing in this study were excluded from the analysis. The database contains 444 (65.0%) benign samples and 239 (35.0%) samples malignas.
comparison and discussion
results obtained in this study were compared with other results, specifically with the current algorithms of greater success, on the basis of data of Wisconsin.
The AMMLP obtained an accuracy in the classification of 99,63% in the best simulation and 99,58% average, improving the results of the rest of classifiers. In addition the AMMLP, in comparison with other algorithms, exhibits a low computational cost and is easy to implement. The success of the system proposed by the researchers of the UPM reinforces some of the hypotheses of Abraham, and establishes new ones, which could lead to important consequences not only in medicine but in psychology and cybernetics.
