Extreme Learning Machine Framework for Risk Stratification of Fatty Liver Disease Using Ultrasound Tissue Characterization

Fatty Liver Disease (FLD) is caused by the deposition of fat in liver cells and leads to deadly diseases such as liver cancer. Several FLD detection and characterization systems using machine learning (ML) based on Support Vector Machines (SVM) have been applied. These ML systems utilize large number of ultrasonic grayscale features, pooling strategy for selecting the best features and several combinations of training/testing. As result, they are computationally intensive, slow and do not guarantee high performance due to mismatch between grayscale features and classifier type. This study proposes a reliable and fast Extreme Learning Machine (ELM)-based tissue characterization system (a class of Symtosis) for risk stratification of ultrasound liver images. ELM is used to train single layer feed forward neural network (SLFFNN). The input-to-hidden layer weights are randomly generated reducing computational cost. The only weights to be trained are hidden-to-output layer which is done in a single pass (without any iteration) making ELM faster than conventional ML methods. Adapting four types of K-fold cross-validation (K = 2, 3, 5 and 10) protocols on three kinds of data sizes: S0-original, S4-four splits, S8-sixty four splits (a total of 12 cases) and 46 types of grayscale features, we stratify the FLD US images using ELM and benchmark against SVM. Using the US liver database of 63 patients (27 normal/36 abnormal), our results demonstrate superior performance of ELM compared to SVM, for all cross-validation protocols (K2, K3, K5 and K10) and all types of US data sets (S0, S4, and S8) in terms of sensitivity, specificity, accuracy and area under the curve (AUC). Using the K10 cross-validation protocol on S8 data set, ELM showed an accuracy of 96.75% compared to 89.01% for SVM, and correspondingly, the AUC: 0.97 and 0.91, respectively. Further experiments also showed the mean reliability of 99% for ELM classifier, along with the mean speed improvement of 40% using ELM against SVM. We validated the symtosis system using two class biometric facial public data demonstrating an accuracy of 100%.     

Prediction of Software Reliability using Bio Inspired Soft Computing Techniques

A lot of models have been made for predicting software reliability. The reliability models are restricted to using particular types of methodologies and restricted number of parameters. There are a number of techniques and methodologies that may be used for reliability prediction. There is need to focus on parameters consideration while estimating reliability. The reliability of a system may increase or decreases depending on the selection of different parameters used. Thus there is need to identify factors that heavily affecting the reliability of the system. In present days, reusability is mostly used in the various area of research. Reusability is the basis of Component-Based System (CBS). The cost, time and human skill can be saved using Component-Based Software Engineering (CBSE) concepts. CBSE metrics may be used to assess those techniques which are more suitable for estimating system reliability. Soft computing is used for small as well as large-scale problems where it is difficult to find accurate results due to uncertainty or randomness. Several possibilities are available to apply soft computing techniques in medicine related problems. Clinical science of medicine using fuzzy-logic, neural network methodology significantly while basic science of medicine using neural-networks-genetic algorithm most frequently and preferably. There is unavoidable interest shown by medical scientists to use the various soft computing methodologies in genetics, physiology, radiology, cardiology and neurology discipline. CBSE boost users to reuse the past and existing software for making new products to provide quality with a saving of time, memory space, and money. This paper focused on assessment of commonly used soft computing technique like Genetic Algorithm (GA), Neural-Network (NN), Fuzzy Logic, Support Vector Machine (SVM), Ant Colony Optimization (ACO), Particle Swarm Optimization (PSO), and Artificial Bee Colony (ABC). This paper presents working of soft computing techniques and assessment of soft computing techniques to predict reliability. The parameter considered while estimating and prediction of reliability are also discussed. This study can be used in estimation and prediction of the reliability of various instruments used in the medical system, software engineering, computer engineering and mechanical engineering also. These concepts can be applied to both software and hardware, to predict the reliability using CBSE.     

任务教学法在医学硕士研究生英语口语教学中的实践及教学体会

本文探讨任务教学法在研究生医学英语口语教学的初步应用.先介绍任务教学法的概念,指出其可以为英语口语教学带来进步.具体描述按照任务教学法采取的教学活动,学生们非常欢迎并积极参与.讨论教学中遇到的问题,并介绍教学体会.最后指出教师和学生应共同努力,借鉴其他口语教学模式,丰富任务教学法,在全面提高医学英语水平的同时,真正提高医学英语口语.     

Avoiding diagnostic pitfalls in mimics of neoplasia: the importance of a comprehensive diagnostic approach

Any medical diagnosis should take a multimodal approach, especially those involving tumour-like conditions, as entities that mimic neoplasms have overlapping features and may present detrimental outcomes if they are underdiagnosed. These case reports present diagnostic pitfalls resulting from overdependence on a single diagnostic parameter for three musculoskeletal neoplasm mimics: brown tumour (BT) that was mistaken for giant cell tumour (GCT), methicillin-resistant Staphylococcus aureus osteomyelitis mistaken for osteosarcoma and a pseudoaneurysm mistaken for a soft tissue sarcoma. Literature reviews revealed five reports of BT simulating GCT, four reports of osteomyelitis mimicking osteosarcoma and five reports of a pseudoaneurysm imitating a soft tissue sarcoma. Our findings highlight the therapeutic dilemmas that arise with musculoskeletal mimics, as well as the importance of thorough investigation to distinguish mimickers from true neoplasms.     

Extracting Information from Electronic Medical Records to Identify the Obesity Status of a Patient Based on Comorbidities and Bodyweight Measures

Obesity is a chronic disease with an increasing impact on the world’s population. In this work, we present a method of identifying obesity automatically using text mining techniques and information related to body weight measures and obesity comorbidities. We used a dataset of 3015 de-identified medical records that contain labels for two classification problems. The first classification problem distinguishes between obesity, overweight, normal weight, and underweight. The second classification problem differentiates between obesity types: super obesity, morbid obesity, severe obesity and moderate obesity. We used a Bag of Words approach to represent the records together with unigram and bigram representations of the features. We implemented two approaches: a hierarchical method and a nonhierarchical one. We used Support Vector Machine and Naïve Bayes together with ten-fold cross validation to evaluate and compare performances. Our results indicate that the hierarchical approach does not work as well as the nonhierarchical one. In general, our results show that Support Vector Machine obtains better performances than Naïve Bayes for both classification problems. We also observed that bigram representation improves performance compared with unigram representation.     

Machine learning–based heart disease prediction system for Indian population: An exploratory study done in South India

BackgroundIn India, huge mortality occurs due to cardiovascular diseases (CVDs) as these diseases are not diagnosed in early stages. Machine learning (ML) algorithms can be used to build efficient and economical prediction system for early diagnosis of CVDs in India.MethodsA total of 1670 anonymized medical records were collected from a tertiary hospital in South India. Seventy percent of the collected data were used to train the prediction system. Five state-of-the-art ML algorithms (k-Nearest Neighbours, Naïve Bayes, Logistic Regression, AdaBoost and Random Forest [RF]) were applied using Python programming language to develop the prediction system. The performance was evaluated over remaining 30% of data. The prediction system was later deployed in the cloud for easy accessibility via Internet.ResultsML effectively predicted the risk of heart disease. The best performing (RF) prediction system correctly classified 470 out of 501 medical records thus attaining a diagnostic accuracy of 93.8%. Sensitivity and specificity were observed to be 92.8% and 94.6%, respectively. The prediction system attained positive predictive value of 94% and negative predictive value of 93.6%. The prediction model developed in this study can be accessed at http://das.southeastasia.cloudapp.azure.com/predict/ConclusionsML-based prediction system developed in this study performs well in early diagnosis of CVDs and can be accessed via Internet. This study offers promising results suggesting potential use of ML-based heart disease prediction system as a screening tool to diagnose heart diseases in primary healthcare centres in India, which would otherwise get undetected.     

极端学习机模型在张家口市手足口病发病率预测中的应用

摘要：目的 探讨极端学习机模型模型在手足口病发病率预测中的应用,并与神经网络模型进行比较。方法 收集2008年5月至017年7月张家口市手足口病月发病率资料,并组成具有111个数据的时间序列,随机选择数据集中75%的数据进行学习建模,剩余25%数据作为预测的检验数据,并对两种模型的预测效果进行验证。结果 ELM学习的MRE为0.05,预测的MRE为：0.07,神经网络学习的MRE为：0.09,预测的MRE为0.12。结论 ELM模型学习效果和预测效果优于神经网络,它可以提高预测的精度,具有较高的实用价值。     

Integrated approach to e-learning enhanced both subjective and objective knowledge of aEEG in a neonatal intensive care unit

INTRODUCTION

Various meta-analyses have shown that e-learning is as effective as traditional methods of continuing professional education. However, there are some disadvantages to e-learning, such as possible technical problems, the need for greater self-discipline, cost involved in developing programmes and limited direct interaction. Currently, most strategies for teaching amplitude-integrated electroencephalography (aEEG) in neonatal intensive care units (NICUs) worldwide depend on traditional teaching methods.

METHODS

We implemented a programme that utilised an integrated approach to e-learning. The programme consisted of three sessions of supervised protected time e-learning in an NICU. The objective and subjective effectiveness of the approach was assessed through surveys administered to participants before and after the programme.

RESULTS

A total of 37 NICU staff (32 nurses and 5 doctors) participated in the study. 93.1% of the participants appreciated the need to acquire knowledge of aEEG. We also saw a statistically significant improvement in the subjective knowledge score (p = 0.041) of the participants. The passing rates for identifying abnormal aEEG tracings (defined as ≥ 3 correct answers out of 5) also showed a statistically significant improvement (from 13.6% to 81.8%, p < 0.001). Among the participants who completed the survey, 96.0% felt the teaching was well structured, 77.8% felt the duration was optimal, 80.0% felt that they had learnt how to systematically interpret aEEGs, and 70.4% felt that they could interpret normal aEEG with confidence.

CONCLUSION

An integrated approach to e-learning can help improve subjective and objective knowledge of aEEG.     

Skin prick test--more reliable than estimation of specific IgE in allergy diagnosis

The aim of this study is to determine the reliability of specific IgE levels in the diagnosis of allergic disorders when compared with the skin prick test. The modified skin prick test has been used in "allergy" diagnosis for several decades in clinics around the world. In recent years however, its position as the numero uno investigation has been challenged by estimation of specific IgE levels to individual allergens. In this study the sensitivity/reliability of "specific IgE" test using the "immunoblot" strip method was compared with the skin prick test. One hundred patients suffering from various allergic disorders underwent skin prick tests and specific IgE for 3 different mites. The results of this study show that the skin prick test is more sensitive/reliable as compared to estimation of specific IgE levels.     

基于深度学习的疾病诊断

以UCI心脏病数据集为试验测试数据,利用机器学习中的深度学习技术进行疾病诊断分类,分别与随机森林、支持向量机以及神经网络分类进行对比,指出深度学习技术对医疗大数据的挖掘具有巨大潜力和重要的应用价值。