Improved Alzheimer’s Disease Detection by MRI Using Multimodal Machine Learning Algorithms

Adult-onset dementia disorders represent a challenge for modern medicine. Alzheimer’s disease (AD) represents the most diffused form of adult-onset dementias. For half a century, the diagnosis of AD was based on clinical and exclusion criteria, with an accuracy of 85%, which did not allow for a definitive diagnosis, which could only be confirmed by post-mortem evaluation. Machine learning research applied to Magnetic Resonance Imaging (MRI) techniques can contribute to a faster diagnosis of AD and may contribute to predicting the evolution of the disease. It was also possible to predict individual dementia of older adults with AD screening data and ML classifiers. To predict the AD subject status, the MRI demographic information and pre-existing conditions of the patient can help to enhance the classifier performance. In this work, we proposed a framework based on supervised learning classifiers in the dementia subject categorization as either AD or non-AD based on longitudinal brain MRI features. Six different supervised classifiers are incorporated for the classification of AD subjects and results mentioned that the gradient boosting algorithm outperforms other models with 97.58% of accuracy.

A Comparative Study of Detection and Classification of Emotions on Social Media Using SVM and Näıve Bayes Techniques

The rapid rate of innovations and dynamics of technology has made humans life more dependent on them. In today’s synopsis Microblogging and Social networking sites like Twitter, Facebook are a part of our lives that cannot be detached from anyone. Through these social media each one of them carry their emotions and fix their opinions based on a particular situations or circumstances. This paper presents a brief comparison about Detection and Classification of Emotions on Social Media using SVM and Näıve Bayesian classifier. Twitter messages has been used as input dataset because they contain a broad, varied, and freely accessible set of emotions. The approach uses hash-tags as labels to train supervised classifiers to detect multiple classes of emotion on potentially large data sets without the need for manual intervention. We look into the usefulness of a number of features for detecting emotions, including unigrams, unigram symbol, negations and punctuations using SVM and Näıve Bayesian Classifiers.

Report Classification for Semiconductor Failure Analysis

In their daily work, engineers in the Failure Analysis (FA) laboratory generate numerous documents reporting all their tasks, findings, and conclusions regarding every device they are handled. This data stores valuable knowledge for the laboratory that other experts can consult, however, the nature of it, as individual reports reporting concrete devices and their corresponding processes, makes it inefficient to consult for the human experts. In this context, the following paper proposes a Artificial Intelligence solution for the gathering of this FA knowledge stored in the numerous documents generated in the laboratory. Therefore, we have generated a dataset of FA reports along with their corresponding electrical signatures and physical failures in order to train different supervised classifiers. The results show that the models are able of capturing the patterns underlying the different jobs and predict the causes, showing slightly better results for the physical hypotheses.

Status and Trends of Wetland Studies in Canada Using Remote Sensing Technology with a Focus on Wetland Classification: A Bibliographic Analysis

A large portion of Canada is covered by wetlands; mapping and monitoring them is of great importance for various applications. In this regard, Remote Sensing (RS) technology has been widely employed for wetland studies in Canada over the past 45 years. This study evaluates meta-data to investigate the status and trends of wetland studies in Canada using RS technology by reviewing the scientific papers published between 1976 and the end of 2020 (300 papers in total). Initially, a meta-analysis was conducted to analyze the status of RS-based wetland studies in terms of the wetland classification systems, methods, classes, RS data usage, publication details (e.g., authors, keywords, citations, and publications time), geographic information, and level of classification accuracies. The deep systematic review of 128 peer-reviewed articles illustrated the rising trend in using multi-source RS datasets along with advanced machine learning algorithms for wetland mapping in Canada. It was also observed that most of the studies were implemented over the province of Ontario. Pixel-based supervised classifiers were the most popular wetland classification algorithms. This review summarizes different RS systems and methodologies for wetland mapping in Canada to outline how RS has been utilized for the generation of wetland inventories. The results of this review paper provide the current state-of-the-art methods and datasets for wetland studies in Canada and will provide direction for future wetland mapping research.

Dwell Time Estimation of Import Containers as an Ordinal Regression Problem

The optimal stacking of import containers in a terminal reduces the reshuffles during the unloading operations. Knowing the departure date of each container is critical for optimal stacking. However, such a date is rarely known because it depends on various attributes. Therefore, some authors have proposed estimation algorithms using supervised classification. Although supervised classifiers can estimate this dwell time, the variable “dwell time” takes ordered values for this problem, suggesting using ordinal regression algorithms. Thus, we have compared an ordinal regression algorithm (selected from 15) against two supervised classifiers (selected from 30). We have set up two datasets with data collected in a container terminal. We have extracted and evaluated 35 attributes related to the dwell time. Additionally, we have run 21 experiments to evaluate both approaches regarding the mean absolute error modified and the reshuffles. As a result, we have found that the ordinal regression algorithm outperforms the supervised classifiers, reaching the lowest mean absolute error modified in 15 (71%) and the lowest reshuffles in 14 (67%) experiments.

Smartphone-Based Human Sitting Behaviors Recognition Using Inertial Sensor

At present, people spend most of their time in passive rather than active mode. Sitting with computers for a long time may lead to unhealthy conditions like shoulder pain, numbness, headache, etc. To overcome this problem, human posture should be changed for particular intervals of time. This paper deals with using an inertial sensor built in the smartphone and can be used to overcome the unhealthy human sitting behaviors (HSBs) of the office worker. To monitor, six volunteers are considered within the age band of 26 ± 3 years, out of which four were male and two were female. Here, the inertial sensor is attached to the rear upper trunk of the body, and a dataset is generated for five different activities performed by the subjects while sitting in the chair in the office. Correlation-based feature selection (CFS) technique and particle swarm optimization (PSO) methods are jointly used to select feature vectors. The optimized features are fed to machine learning supervised classifiers such as naive Bayes, SVM, and KNN for recognition. Finally, the SVM classifier achieved 99.90% overall accuracy for different human sitting behaviors using an accelerometer, gyroscope, and magnetometer sensors.

Melanoma Detection from Lesion Images using Optimized Features Selected by Metaheuristic Algorithms

This paper deals with a simple but efficient method for detection of deadly malignant melanoma with optimized hand-crafted feature sets selected by three alternative metaheuristic algorithms, namely Particle Swarm Optimization (PSO), Ant Colony Optimization (ACO) and Simulated Annealing (SA). Total 1898 number of features relating to lesion shapes, colors and textures are extracted from each of the 170 non-dermoscopy camera images of the popular MED-NODE dataset. This large feature set is then optimized and the number of features is reduced to up-to the range of single digit using metaheuristic algorithms as feature selector. Two well-known supervised classifiers, i.e. Support Vector Machine (SVM) and Artificial Neural Network (ANN) are used to classify malignant and benign lesions. The best classification accuracy result found by this method is 87.69% with only 7 features selected by PSO using ANN classifier which is far better than the results found in the literature so far.

Dynamic Nearest Neighbor: An Improved Machine Learning Classifier and Its Application in Finances

The presence of machine learning, data mining and related disciplines is increasingly evident in everyday environments. The support for the applications of learning techniques in topics related to economic risk assessment, among other financial topics of interest, is relevant for us as human beings. The content of this paper consists of a proposal of a new supervised learning algorithm and its application in real world datasets related to finance, called D1-NN (Dynamic 1-Nearest Neighbor). The D1-NN performance is competitive against the main state of the art algorithms in solving finance-related problems. The effectiveness of the new D1-NN classifier was compared against five supervised classifiers of the most important approaches (Bayes, nearest neighbors, support vector machines, classifier ensembles, and neural networks), with superior results overall.

A Novel Hierarchical Framework for Measuring the Complexity and Irregularity of Multimodal Speech Signals and Its Application in the Assessment of Speech Impairment in Amyotrophic Lateral Sclerosis

Purpose The purposes of this study are to develop a novel multimodal framework for measuring variability at the muscular, kinematic, and acoustic levels of the motor speech hierarchy and evaluate the utility of this framework in detecting speech impairment in amyotrophic lateral sclerosis (ALS). Method The myoelectric activities of three bilateral jaw muscle pairs (masseter, anterior temporalis, and anterior belly of digastric), jaw kinematics, and speech acoustics were recorded in 13 individuals with ALS and 10 neurologically healthy controls during sentence reading. Thirteen novel measures (six muscular, three kinematic, four acoustic), which characterized two different but interrelated aspects of variability—complexity and irregularity—were derived using linear and nonlinear methods. Exploratory factor analysis was applied to identify the latent factors underlying these measures. Based on the latent factors, three supervised classifiers—support vector machine (SVM), random forest (RF), and logistic regression (Logit)—were used to differentiate between the speech samples for patients and controls. Results Four interpretable latent factors were identified, representing the complexity of jaw kinematics, the irregularity of jaw antagonists functioning, the irregularity of jaw agonists functioning, and the irregularity of subband acoustic signals, respectively. Based on these latent factors, the speech samples for patients and controls were classified with high accuracy (> 96% for SVM and RF; 88.64% for Logit), outperforming the unimodal measures. Two factors showed significant between-groups differences, as characterized by decreased complexity of jaw kinematics and increased irregularity of jaw antagonists functioning in patients versus controls. Conclusions Decreased complexity of jaw kinematics presumably reflects impaired fine control of jaw movement, while increased irregularity of jaw antagonists functioning could be attributed to reduced synchronization of motor unit firing in ALS. The findings provide preliminary evidence for the utility of the multimodal framework as a novel quantitative assessment tool for detecting speech impairment in ALS and (potentially) in other neuromotor disorders.

Analyzing Linguistic Features for Classifying Why-Type Non-Factoid Questions

Why-type non-factoid questions are complex and difficult to answer compared to factoid questions. A challenge in finding an accurate answer to a non-factoid question is to understand the intent of user as it differs with their knowledge and also the context of the question in which it is being asked. Predicting correct type of a question and its answer by a classification model is an important issue as it affects the subsequent processing of its answer. In this paper, a classification model is proposed which is trained by a combination of lexical, syntactic, and semantic features to classify open-domain why-type questions. Various supervised classifiers are trained on a featured dataset out of which support vector machine achieves the highest accuracy of 81% in determining question type and 76.8% in determining answer type which shows 14.6% improvement in predicting an answer type over a baseline why-type classifier with 62.2% accuracy.