Machine Learning Software Research Articles

Use of automated artificial intelligence to predict the need for orthodontic extractions.

ObjectiveTo develop and explore the usefulness of an artificial intelligence system for the prediction of the need for dental extractions during orthodontic treatments based on gender, model variables, and cephalometric records.MethodsThe gender, model variables, and radiographic records of 214 patients were obtained from an anonymized data bank containing 314 cases treated by two experienced orthodontists. The data were processed using an automated machine learning software (Auto-WEKA) and used to predict the need for extractions.ResultsBy generating and comparing several prediction models, an accuracy of 93.9% was achieved for determining whether extraction is required or not based on the model and radiographic data. When only model variables were used, an accuracy of 87.4% was attained, whereas a 72.7% accuracy was achieved if only cephalometric information was used.ConclusionsThe use of an automated machine learning system allows the generation of orthodontic extraction prediction models. The accuracy of the optimal extraction prediction models increases with the combination of model and cephalometric data for the analytical process.

Application of Machine Learning in the Reliability Evaluation of Pipelines for the External Anticorrosion Coating.

The purpose of this research is to enhance the analysis of the reliability status for external anticorrosive coatings. With the limitation and insufficiency of the static evaluation method, we study and construct an evaluation method of dynamic reliability for the anticorrosive layer, integrating the trend analysis of the Markov chain and the set pair theory. This method is implemented by the machine learning software of PyCharm community edition, based on Python language. The algorithm utilizes the connection degree in the set pair theory to determine the risk levels of the anticorrosive coating systems. According to the characteristics of the dynamic change of the anticorrosive layer with time, we built the mathematical evaluation model by combining it with the nonaftereffect property of the Markov chain. Therefore, we can make a dynamic and useful analysis for the reliability grade of the anticorrosive coating and assess the effectiveness grade of the changed reliability for the anticorrosive coating after some time. This method can effectively evaluate the reliability level of the anticorrosion coating through the example of big data of detection points. Under national standards, we provide the theoretical basis for pipeline maintenance within detection cycle requirements.

Dental age assessment based on CBCT images using machine learning algorithms

Age estimation has become inordinately significant for human beings for many reasons, such as detecting legal and criminal responsibility and other social events like a marriage license, birth certificate, etc. This paper aims to decide on the most desirable machine learning algorithm (from conventional machine learning algorithms to deep learning) for dental age estimation based on buccal bone level. The database consisted of 150 CBCT images (73 males and 77 females) from an existing base of the Faculty of Dental Medicine with Clinics, University of Sarajevo, aged 20–69. Results were obtained using the Waikato Environment for Knowledge Analysis (Weka), machine learning software in Java. Left and Right Buccal Alveolar Bone Levels are increasing with age, so they showed to be the most important attributes, especially the latter. Random Forest classifier provided the greatest result with the correlation coefficient of 0.803 and the mean absolute error of 6.022. We have also shown that considering sinus-related features can be a significant addition to the databases. Our paper is probably one of the first studies where regression algorithms based on the Support Vector Machines and Random Forest were utilized.

Towards an AI-friendly cross-timescale simulation and analysis platform for electric distribution systems

Substantial changes are occurring in electric distribution systems due to ambitious targets towards carbon-neutrality in many regions around the world. One of the key challenges is how to analyze the interactions of massive amount of energy end-users with the electric distribution grid operator. In this paper, we introduce a comprehensive simulation platform, Al4Dist, that is capable to perform a wide collection of distribution system studies that capture multiple timescales ranging from market planning to transient event analysis. Al4Dist is designed to effortlessly integrate with off-the-shelf machine learning packages and algorithm implementations. We envision that Al4Dist will serve as a platform to empower researchers with different expertise to contribute to the development of low carbon electricity sector.

How the Brunswikian Lens Model Illustrates the Relationship Between Physiological and Behavioral Signals and Psychological Emotional and Cognitive States.

Social relationships are constructed by and through the relational communication that people exchange. Relational messages are implicit nonverbal and verbal messages that signal how people regard one another and define their interpersonal relationships—equal or unequal, affectionate or hostile, inclusive or exclusive, similar or dissimilar, and so forth. Such signals can be measured automatically by the latest machine learning software tools and combined into meaningful factors that represent the socioemotional expressions that constitute relational messages between people. Relational messages operate continuously on a parallel track with verbal communication, implicitly telling interactants the current state of their relationship and how to interpret the verbal messages being exchanged. We report an investigation that explored how group members signal these implicit messages through multimodal behaviors measured by sensor data and linked to the socioemotional cognitions interpreted as relational messages. By use of a modified Brunswikian lens model, we predicted perceived relational messages of dominance, affection, involvement, composure, similarity and trust from automatically measured kinesic, vocalic and linguistic indicators. The relational messages in turn predicted the veracity of group members. The Brunswikian Lens Model offers a way to connect objective behaviors exhibited by social actors to the emotions and cognitions being perceived by other interactants and linking those perceptions to social outcomes. This method can be used to ascertain what behaviors and/or perceptions are associated with judgments of an actor’s veracity. Computerized measurements of behaviors and perceptions can replace manual measurements, significantly expediting analysis and drilling down to micro-level measurement in a previously unavailable manner.

COVID-19 Face Mask Detection

Abstract: Face mask detection involves in detection the placement of the face then crucial whether or not it's a mask thereon or not. the problem is proximately cognate to general object notion to detect the categories of objects. Face identification flatly deals with identifying a particular cluster of entities i.e., Face. it's varied applications, like autonomous driving, education, police work, and so on. This paper presents a simplified approach to serve the above purpose using the basic Machine Learning (ML) packages such as TensorFlow, Keras, OpenCV and Scikit-Learn. The planned technique detects the face from the image properly and so identifies if it's a mask on that or not. As an investigation taskperforming artist, it ought to conjointly sight a face at the side of a mask in motion. The technique perform accuracyup to 95.77% and 94.58% respectively on two different datasets and count optimized values of parameters using the Sequential Convolutional Neural Network model to detect the presence of masks correctly without causing over-fitting. Keywords: TensorFlow, Keras, OpenCV and Scikit- Learn

Hill Climb Game Play with Webcam Using OpenCV

Abstract: Computer vision is a process by which we can understand how the images and videos are stored and manipulated, also it helps in the process of retrieving data from either images or videos. Computer Vision is part of Artificial Intelligence. Computer-Vision plays a major role in Autonomous cars, Object detections, robotics, object tracking, etc. OpenCV (Open Source Computer Vision Library) is an open source computer vision and machine learning software library. OpenCV was built to provide a common infrastructure for computer vision applications and to accelerate the use of machine perception in the commercial products. It comes with a highly improved deep learning (dnn ) module. This module now supports a number of deep learning frameworks, including Caffe, TensorFlow, and Torch/PyTorch. This does allow us to take our models trained using dedicated deep learning libraries/tools and then efficiently use them directly inside our OpenCV scripts. MediaPipe is a framework mainly used for building audio, video, or any time series data. With the help of the MediaPipe framework, we can build very impressive pipelines for different media processing functions like Multi-hand Tracking, Face Detection, Object Detection and Tracking, etc.

MitoDataclean: A machine learning approach for the accurate identification of cross-contamination-derived tumor mitochondrial DNA mutations.

Next-generation sequencing (NGS) of mitochondrial DNA (mtDNA) has widespread applications in aging and cancer studies. However, cross-contamination of mtDNA constitutes a major concern. Previous methods for the detection of mtDNA contamination mainly focus on haplogroup-level phylogeny, but neglect haplotype-level differences, leading to limited sensitivity and accuracy. In our study, we present mitoDataclean, a random-forest-based machine learning package for accurate identification of cross-contamination, evaluation of contamination levels and detection of contamination-derived variants in mtDNA NGS data. Comprehensive optimization of mitoDataclean revealed that training simulation with mixtures of small haplogroup distance and low polymorphic difference was critical for optimal modeling. Compared to existing methods, mitoDataclean exhibited significantly improved sensitivity and accuracy for the detection of sample contamination in simulated data. In addition, mitoDataclean achieved area under the curve values of 0.91 and 0.97 for discerning genuine and contamination-derived mtDNA variants in a simulated Western dataset and private sequencing contamination data, respectively, suggesting that this tool may be applicable for different populations and samples with different sources of contamination. Finally, mitoDataclean was further evaluated in several private and public datasets and showed a robust ability for contamination detection. Altogether, our study demonstrates that mitoDataclean may be used for accurate detection of contaminated samples and contamination-derived variants in mtDNA NGS data.

Smoke testing for machine learning: simple tests to discover severe bugs

Machine learning is nowadays a standard technique for data analysis within software applications. Software engineers need quality assurance techniques that are suitable for these new kinds of systems. Within this article, we discuss the question whether standard software testing techniques that have been part of textbooks since decades are also useful for the testing of machine learning software. Concretely, we try to determine generic and simple smoke tests that can be used to assert that basic functions can be executed without crashing. We found that we can derive such tests using techniques similar to equivalence classes and boundary value analysis. Moreover, we found that these concepts can also be applied to hyperparameters, to further improve the quality of the smoke tests. Even though our approach is almost trivial, we were able to find bugs in all three machine learning libraries that we tested and severe bugs in two of the three libraries. This demonstrates that common software testing techniques are still valid in the age of machine learning and that considerations how they can be adapted to this new context can help to find and prevent severe bugs, evenin mature machine learning libraries.

Pretreatment Contrast-Enhanced Computed Tomography Radiomics for Prediction of Pathological Regression Following Neoadjuvant Chemotherapy in Locally Advanced Gastric Cancer: A Preliminary Multicenter Study

BackgroundSensitivity to neoadjuvant chemotherapy in locally advanced gastric cancer patients varies; however, an effective predictive marker is currently lacking. We aimed to propose and validate a practical treatment efficacy prediction method based on contrast-enhanced computed tomography (CECT) radiomics.MethodData of l24 locally advanced gastric carcinoma patients who underwent neoadjuvant chemotherapy were acquired retrospectively between December 2012 and August 2020 from three different cancer centers. In total, 1216 radiomics features were initially extracted from each lesion’s pretreatment portal venous phase computed tomography image. Subsequently, a radiomics predictive model was constructed using machine learning software. Clinicopathological data and radiological parameters of the enrolled patients were collected and analyzed retrospectively. Univariate and multivariate logistic regression analyses were performed to screen for independent predictive indices. Finally, we developed an integrated model combining clinicopathological predictive parameters and radiomics features.ResultIn the training set, 10 (14.9%) patients achieved a good response (GR) after preoperative neoadjuvant chemotherapy (n = 77), whereas in the testing set, seven (17.5%) patients achieved a GR (n = 47). The radiomics predictive model showed competitive prediction efficacy in both the training and independent external validation sets. The areas under the curve (AUC) values were 0.827 (95% confidence interval [CI]: 0.609–1.000) and 0.854 (95% CI: 0.610–1.000), respectively. Similarly, when only the single hospital data were included as an independent external validation set (testing set 2), AUC values of the models were 0.827 (95% CI: 0.650–0.952) and 0.889 (95% CI: 0.663–1.000) in the training set and testing set 2, respectively.ConclusionOur study is the first to discover that CECT radiomics could provide powerful and consistent predictions of therapeutic sensitivity to neoadjuvant chemotherapy among gastric cancer patients across different hospitals.

FISSURE IDENTIFICATION METHODS IN RICE SEEDS AFTER ARTIFICIAL DRYING

New, efficient, low-cost techniques for image processing and alternative machine learning for seed processing are of academic and industrial interest. This study aims to identify fissures in bark and peeled rice seeds using X-ray and RGB image processing techniques and machine learning. Samples of three batches of rice seeds were used: a batch of seeds not subjected to drying (peeled seed), and the other two comprised of dried seeds, one containing seeds with husk and another containing huskless seeds; each sample comprised 100 seeds. Images in X-ray and RGB formats were provided in the sequence processed in ImageJ software and introduced in the machine learning software, where they were pre-processed using the appropriate filters and then classified by the J48 and linear discriminant analysis (LDA) classifiers. X-ray images obtained using differentiated equipment allow the identification of cracks in rice seeds using image processing techniques and the LDA classifier. Capturing images using RGB is a viable alternative. Using filters, either individually or in combination, may constitute an adequate alternative for rice seed classification.

A Genetic Algorithm and PCA-Based Feature Selection to Improve the Failure Diagnosis Performance of Railway Vehicle Doors

The failure diagnosis of railway vehicle door system is carried out using a test bench and machine learning software for the fast and accurate classification. The signal length deviation exists in actual collected data of normal operation and abnormal failures with a time delay. The traditional data multi-segmentation technique for feature extraction has shortcomings by assuming that the measured time-based signals have the same operating time. However, the uniform data segmentation has a difficulty due to the deviation of measured data length. A method of converting time-based data into position-based data was performed to overcome the deviation problem. A method of optimized single-zone data using a genetic algorithm was proposed to improve the classification performance and to reduce computation time, instead of existing the multi-segmentation technique. A principal component analysis-based feature dimensional reduction with explained variance ratio was used to reduce the effect from multi-collinearity of features. Finally, the combination of the proposed methods was compared with individual methods to validate the classification performance by using support vector machine and other classifiers. It was confirmed that the proposed combination method shows the highest classification accuracy of 99.84%.

Probabilistic machine learning for breast cancer classification.

A probabilistic neural network has been implemented to predict the malignancy of breast cancer cells, based on a data set, the features of which are used for the formulation and training of a model for a binary classification problem. The focus is placed on considerations when building the model, in order to achieve not only accuracy but also a safe quantification of the expected uncertainty of the calculated network parameters and the medical prognosis. The source code is included to make the results reproducible, also in accordance with the latest trending in machine learning research, named Papers with Code. The various steps taken for the code development are introduced in detail but also the results are visually displayed and critically analyzed also in the sense of explainable artificial intelligence. In statistical-classification problems, the decision boundary is the region of the problem space in which the classification label of the classifier is ambiguous. Problem aspects and model parameters which influence the decision boundary are a special aspect of practical investigation considered in this work. Classification results issued by technically transparent machine learning software can inspire more confidence, as regards their trustworthiness which is very important, especially in the case of medical prognosis. Furthermore, transparency allows the user to adapt models and learning processes to the specific needs of a problem and has a boosting influence on the development of new methods in relevant machine learning fields (transfer learning).

Developing and Implementing Machine Learning Software at Home Depot

This teaching case explores the problem of shelfouts and the use of technology adoption to minimize it. Shelfout, wherein a product is not on the shelf when it is supposed to be, has received renewed interest especially given the fact that many brick-and-mortar stores shut down due to their inability to compete with online vendors. The coronavirus pandemic worsened this problem and companies continue to struggle with the resulting supply chain disruptions. Increasingly consumers are searching for products on the website to confirm product availability before traveling to the store. In this case we show how The Home Depot, Inc., (Home Depot), is addressing shelfouts and the process they went through in solving this problem. Instructors can use this case to introduce students to the machine learning development lifecycle, marketing courses discussing shelfouts, and courses with lessons related to technology implementation.

DECODING VIEWS AND SENTIMENTS OF PROGRAM HEADS TOWARDS THE SUPERVISION OF INSTRUCTION DURING THE COVID-19 PANDEMIC

The purpose of this study was to decode the hidden views and sentiments from the collated written responses of Eastern Samar State University’s Program Heads regarding supervision of instructions amidst the COVID-19 pandemic. This study utilized Exploratory Sequential Mixed Method to explore and understand the perspective or sentiments of Eastern Samar State University program heads towards supervision of instruction in the midst of the COVID-19 pandemic. Data were collected/collated from the participants indirectly using an interview questionnaire containing an open-ended question. The same were processed and analyzed using an open-source machine learning software called Orange toolbox (Demsar et al., 2013) wherein pre-processing, sentiment analysis and topic modelling built-in tools were utilized. The results showed that the most prominent words generated by the machine learning tool from the text file of responses are the words pandemic, performance, program, learning, difficult, supervision, instruction, internet, faculty, online students, teaching, delivery confusing, challenging, poor and connectivity. The dominant sentiment associated thereof lean towards negative polarity which implicate negative sentiments. Hidden topics were automatically generated by the machine which allowed the researchers to come up with the following related themes: “Impact of pandemic in the supervision of instruction of faculty and learning of students”, “Challenges in the delivery of instruction and supervision due to poor internet connectivity”, and “Strategic role of online modalities and connectivity in supervision and delivery of instruction”. There are limited researches navigating in text mining and sentiment analysis with the use of Orange toolbox particularly those that deals with supervision of instruction in a Philippine State University. There are related studies using machine learning software, but nothing like this study directed towards a specific gap in specific locale. KEYWORDS: Pandemic, Latent Semantic Indexing, Orange Toolbox, Sentiment Analysis, Thematic Analysis.

Improved Classification Accuracy for Identification of Cervical Cancer

The major purpose of this research is to forecast cervical cancer, compare which algorithms perform well, and then choose the best algorithm to predict cervical cancer at an early stage. Cervical cancer classification can be automated using a machine learning system. This study evaluates multiple machine learning techniques for cervical cancer classification. For this classification, algorithms such as Decision Tree, Naive Bayes, KNN, SVM, and MLP are proposed and evaluated. The cervical cancer Dataset, which was retrieved from the UCI machine learning data repository, was used to test these methods. With the help of Sciklit-learn, the algorithms' results were compared in terms of Accuracy, Sensitivity, and Specificity. Sciklit-learn is a Python-based machine learning package that is available for free. Finally, the best model for predicting cervical cancer is developed.

Overview of Imaging Modalities in Stroke.

This article reviews common imaging modalities used in diagnosis and management of acute stroke. Each modality is discussed individually and clinical scenarios are presented to demonstrate how to apply these modalities in decision-making. Advances in neuroimaging provide unprecedented accuracy in determining tissue viability as well as tissue fate in acute stroke. In addition, advances in machine learning have led to the creation of decision support tools to improve the interpretability of these studies. Noncontrast head computed tomography (CT) remains the most commonly used initial imaging tool to evaluate stroke. Its exquisite sensitivity for hemorrhage, rapid acquisition, and widespread availability make it the ideal first study. CT angiography (CTA), the most common follow-up study after noncontrast head CT, is used primarily to identify intracranial large vessel occlusions and cervical carotid or vertebral artery disease. CTA is highly sensitive and can improve accuracy of patient selection for endovascular therapy through delineations of ischemic core. CT perfusion is widely used in endovascular therapy trials and benefits from multiple commercially available machine-learning packages that perform automated postprocessing and interpretation. Magnetic resonance imaging (MRI) and magnetic resonance angiography (MRA) can provide valuable insights for outcomes prognostication as well as stroke etiology. Optical coherence tomography (OCT), positron emission tomography (PET), single-photon emission computerized tomography (SPECT) offer similar insights. In the clinical scenarios presented, we demonstrate how multimodal imaging approaches can be tailored to gain mechanistic insights for a range of cerebrovascular pathologies.

A workflow for low-cost automated image analysis of myxomycete spore numbers, size and shape.

Measuring spore size is a standard method for the description of fungal taxa, but in manual microscopic analyses the number of spores that can be measured and information on their morphological traits are typically limited. To overcome this weakness we present a method to analyze the size and shape of large numbers of spherical bodies, such as spores or pollen, by using inexpensive equipment. A spore suspension mounted on a slide is treated with a low-cost, high-vibration device to distribute spores uniformly in a single layer without overlap. Subsequently, 10,000 to 50,000 objects per slide are measured by automated image analysis. The workflow involves (1) slide preparation, (2) automated image acquisition by light microscopy, (3) filtering to separate high-density clusters, (4) image segmentation by applying a machine learning software, Waikato Environment for Knowledge Analysis (WEKA), and (5) statistical evaluation of the results. The technique produced consistent results and compared favorably with manual measurements in terms of precision. Moreover, measuring spore size distribution yields information not obtained by manual microscopic analyses, as shown for the myxomycete Physarum albescens. The exact size distribution of spores revealed irregularities in spore formation resulting from the influence of environmental conditions on spore maturation. A comparison of the spore size distribution within and between sporocarp colonies showed large environmental and likely genetic variation. In addition, the comparison identified specimens with spores roughly twice the normal size. The successful implementation of the presented method for analyzing myxomycete spores also suggests potential for other applications.

Objective and bias-free measures of candidate motivation during job applications

Society suffers from biases and discrimination, a longstanding dilemma that stems from ungrounded, subjective judgments. Especially unequal opportunities in labor remain a persistent challenge, despite the recent inauguration of top-down diplomatic measures. Here we propose a solution by using an objective approach to the measurement of nonverbal behaviors of job candidates that trained for a job assessment. First, we implemented and developed artificial intelligence, computer vision, and unbiased machine learning software to automatically detect facial muscle activity and emotional expressions to predict the candidates’ self-reported motivation levels. The motivation judgments by our model outperformed recruiters’ unreliable, invalid, and sometimes biased judgments. These findings mark the necessity and usefulness of novel, bias-free, and scientific approaches to candidate and employee screening and selection procedures in recruitment and human resources.

Comparison of Gait Speed Reserve, Usual Gait Speed, and Maximum Gait Speed of Adults Aged 50+ in Ireland Using Explainable Machine Learning.

Gait speed is a measure of general fitness. Changing from usual (UGS) to maximum (MGS) gait speed requires coordinated action of many body systems. Gait speed reserve (GSR) is defined as MGS-UGS. From a shortlist of 88 features across five categories including sociodemographic, cognitive, and physiological, we aimed to find and compare the sets of predictors that best describe UGS, MGS, and GSR. For this, we leveraged data from 3,925 adults aged 50+ from Wave 3 of The Irish Longitudinal Study on Ageing (TILDA). Features were selected by a histogram gradient boosting regression-based stepwise feature selection pipeline. Each model's feature importance and input-output relationships were explored using TreeExplainer from the Shapely Additive Explanations explainable machine learning package. The mean (SD) from fivefold cross-validation on training data and the score on test data were 0.38 (0.04) and 0.41 for UGS, 0.45 (0.04) and 0.46 for MGS, and 0.19 (0.02) and 0.21 for GSR. Each model selected features across all categories. Features common to all models were age, grip strength, chair stands time, mean motor reaction time, and height. Exclusive to UGS and MGS were educational attainment, fear of falling, Montreal cognitive assessment errors, and orthostatic intolerance. Exclusive to MGS and GSR were body mass index (BMI), and number of medications. No features were selected exclusively for UGS and GSR. Features unique to UGS were resting-state pulse interval, Center for Epidemiologic Studies Depression Scale (CESD) depression, sit-to-stand difference in diastolic blood pressure, and left visual acuity. Unique to MGS were standard deviation in sustained attention to response task times, resting-state heart rate, smoking status, total heartbeat power during paced breathing, and visual acuity. Unique to GSR were accuracy proportion in a sound-induced flash illusion test, Mini-mental State Examination errors, and number of cardiovascular conditions. No interactions were present in the GSR model. The four features that overall gave the most impactful interactions in the UGS and MGS models were age, chair stands time, grip strength, and BMI. These findings may help provide new insights into the multisystem predictors of gait speed and gait speed reserve in older adults and support a network physiology approach to their study.