Data Mining Practical Machine Learning Tools and Techniques

The use of artificial intelligence, and the deep-learning subtype in particular, has been enabled by the use of labeled big data, along with markedly enhanced computing power and cloud storage, across all sectors. In medicine, this is beginning to have an impact at three levels: for clinicians, predominantly via rapid, accurate image interpretation; for health systems, by improving workflow and the potential for reducing medical errors; and for patients, by enabling them to process their own data to promote health. The current limitations, including bias, privacy and security, and lack of transparency, along with the future directions of these applications will be discussed in this article. Over time, marked improvements in accuracy, productivity, and workflow will likely be actualized, but whether that will be used to improve the patient-doctor relationship or facilitate its erosion remains to be seen.

High-performance medicine: the convergence of human and artificial intelligence

Computational Analysis of Microarray Data

Magnetars are highly magnetized young neutron stars that occasionally produce enormous bursts and flares of X-rays and gamma-rays. Of the approximately thirty magnetars currently known in our Galaxy and Magellanic Clouds, five have exhibited transient radio pulsations. Fast radio bursts (FRBs) are millisecond-duration bursts of radio waves arriving from cosmological distances. Some have been seen to repeat. A leading model for repeating FRBs is that they are extragalactic magnetars, powered by their intense magnetic fields. However, a challenge to this model has been that FRBs must have radio luminosities many orders of magnitude larger than those seen from known Galactic magnetars. Here we report the detection of an extremely intense radio burst from the Galactic magnetar SGR 1935+2154 using the Canadian Hydrogen Intensity Mapping Experiment (CHIME) FRB project. The fluence of this two-component bright radio burst and the estimated distance to SGR 1935+2154 together imply a 400-800 MHz burst energy of $\sim 3 \times 10^{34}$ erg, which is three orders of magnitude brighter than those of any radio-emitting magnetar detected thus far. Such a burst coming from a nearby galaxy would be indistinguishable from a typical FRB. This event thus bridges a large fraction of the radio energy gap between the population of Galactic magnetars and FRBs, strongly supporting the notion that magnetars are the origin of at least some FRBs.

A bright millisecond-duration radio burst from a Galactic magnetar

Computational modeling of chemical and biological systems at atomic resolution is a crucial tool in the chemist’s toolset. The use of computer simulations requires a balance between cost and accuracy: quantum-mechanical methods provide high accuracy but are computationally expensive and scale poorly to large systems, while classical force fields are cheap and scalable, but lack transferability to new systems. Machine learning can be used to achieve the best of both approaches. Here we train a general-purpose neural network potential (ANI-1ccx) that approaches CCSD(T)/CBS accuracy on benchmarks for reaction thermochemistry, isomerization, and drug-like molecular torsions. This is achieved by training a network to DFT data then using transfer learning techniques to retrain on a dataset of gold standard QM calculations (CCSD(T)/CBS) that optimally spans chemical space. The resulting potential is broadly applicable to materials science, biology, and chemistry, and billions of times faster than CCSD(T)/CBS calculations. Computational modelling of chemical systems requires a balance between accuracy and computational cost. Here the authors use transfer learning to develop a general purpose neural network potential that approaches quantum-chemical accuracy for reaction thermochemistry, isomerization, and drug-like molecular torsions.

/pdf/approaching-coupled-cluster-accuracy-with-a-general-purpose-11x407tqmn.pdf

Approaching coupled cluster accuracy with a general-purpose neural network potential through transfer learning.

For past several years, microarray technology has attracted tremendous interest for both scientific community and industry. Recently, the applications of microarrays include gene discovery, disease diagnosis and prognosis, drug discovery, etc. High dimensional data with small sample size is the main problem that generate the application of dimension reduction in microarray data analysis. It is seen that SVM, ANN and NB have recently gained wide popularity for cancer classification problems. An efficient and reliable method of dimension reduction plays an important role to improve the performance of SVM, ANN and NB, when applied for classification of high dimensional microarray data. In this book, we applied different combinations of feature selection / extraction methods, as a novel hybrid dimension reduction method for SVM, ANN and NB classifiers. The obtained results are compared with other popular published dimension reduction methods for SVM, NB and ANN classifiers.

Machine Learning Approach

A novel approach for dimension reduction of microarray

Dimension reduction has become inevitable for pre-processing of high dimensional data. “Gene expression microarray data” is an instance of such high dimensional data. Gene expression microarray data displays the maximum number of genes (features) simultaneously at a molecular level with a very small number of samples. The copious numbers of genes are usually provided to a learning algorithm for producing a complete characterization of the classification task. However, most of the times the majority of the genes are irrelevant or redundant to the learning task. It will deteriorate the learning accuracy and training speed as well as lead to the problem of overfitting. Thus, dimension reduction of microarray data is a crucial preprocessing step for prediction and classification of disease. Various feature selection and feature extraction techniques have been proposed in the literature to identify the genes, that have direct impact on the various machine learning algorithms for classification and eliminate the remaining ones. This paper describes the taxonomy of dimension reduction methods with their characteristics, evaluation criteria, advantages and disadvantages. It also presents a review of numerous dimension reduction approaches for microarray data, mainly those methods that have been proposed over the past few years.

Dimension reduction methods for microarray data: a review

Independent component analysis (ICA) is a powerful concept for reducing the dimension of big data in many applications. It has been used for the feature extraction of microarray gene expression data in numerous works. One of the merits of ICA is that a number of extracted features are always equal to the number of samples. When ICA is applied to microarray data, whenever, it faces the challenges of how to find the best subset of genes (features) from extracted features. To resolve this problem, in this paper, we propose a new (artificial bee colony) ABC-based feature selection approach for microarray data. Our approach is based on two stages: ICA-based extraction approach to reduce the size of data and ABC-based wrapper approach to optimize the reduced feature vectors. To validate our proposed approach, extensive experiments were conducted to compare the performance of ICA + ABC with the results obtained from recently published and other previously suggested methods of gene selection for Naive Bayes (NB) classifier. To compare the performance of the proposed approach with other algorithms, a statistical hypothesis test was employed with six benchmark cancer classification datasets of the microarray. The experimental result shows that the proposed approach demonstrates an improvement over all the algorithms for NB classifier with a certain level of significance.

Novel machine learning approach for classification of high-dimensional microarray data

This paper proposed a new combination of feature selection/extraction approach for Artificial Neural Networks (ANNs) classification of high-dimensional microarray data, which uses an Independent Component Analysis (ICA) as an extraction technique and Artificial Bee Colony (ABC) as an optimisation technique. The study evaluates the performance of the proposed ICA + ABC algorithm by conducting extensive experiments on five-binary and one multi-class gene expression microarray data set and compared the proposed algorithm with ICA and ABC. The proposed method shows superior performance as it achieves the highest classification accuracy along with the lowest average number of selected genes. Furthermore, the present work compares the proposed ICA + ABC algorithm with popular filter techniques and with other similar bio-inspired algorithms with ICA. The experimental results show that the proposed algorithm gives more accurate classification rate for ANN classifier. Therefore, ICA + ABC are a promising approach for solving gene selection and cancer classification problems using microarray data.

C. K. Verma

Papers

Machine Learning Approach

A novel approach for dimension reduction of microarray

Dimension reduction methods for microarray data: a review

Novel machine learning approach for classification of high-dimensional microarray data

Artificial neural network classification of microarray data using new hybrid gene selection method