Predicting the Authenticity of Banknotes Using Supervised Learning

The objective of this study was to assess the ability of four peripheral bone measurement sites to predict low bone mineral density of the lumbar spine, femoral neck, and total hip as defined by the WHO classification, and to evaluate their role as a screening procedure for central dual-energy X-ray absorptiometry (DXA). One thousand three hundred white women patients aged 58.3+/-13.9 years were enrolled. Correlations between the peripheral and central sites were moderate, varying between 0.51 and 0.67. In receiver-operating characteristic (ROC) analysis, the areas under the curve for the peripheral sites revealed a few small, but significant, differences between sites. The optimal cut-point was derived from the ROC data for each site to assess its value as a screening parameter. These proved to be unsatisfactory, because the average number of false positives and false negatives were 18% and 24%, respectively. A proposal is made which entails two cut-points, one for an acceptable percentage of false negatives and the other for an acceptable number of false positives. Patients with t-scores between these cut-points would be referred for central DXA for classification. The result is a substantial decrease in the number of false negatives, i.e., patients who would be candidates for treatment. The contentious issue is what proportion of supplemental DXAs is considered logistically and economically acceptable for any of the peripheral site measurements to be useful.

What Interval of Daily Pain Assessment Is Required to Reliably Diagnose Chronic Pain in Sickle Cell Disease? the Pisces Study

With the rapid development of information technology, which is widely used in all spheres of human life and activity, extremely large amounts of data have been accumulated today. By applying machine learning methods to this data, new practically useful knowledge can be obtained. The main goal of this paper is to study different machine learning methods for solving the classification problem and compare their efficiency and accuracy. A separate task is data pre-processing aimed at solving the problem of sample imbalance, as well as identifying the principal components that will be used to solve the classification problem. For this purpose, an information system for classifying the bankruptcy of a company with specified economic and financial characteristics was researched and developed. The study uses a dataset on the basis of which the efficiency and quality of application of several existing classification algorithms are evaluated. These classifiers are: conventional and linear Support Vector Machine, Extra Trees, Random Forest, Decision Tree, Logistic Regression, Multilayer perceptron Classifier, Gradient Boosting, Naive Bayes Classifier. For data pre-processing, we scaled the data, used the SMOTE method to get rid of the imbalance of the training sample, and performed principal component analysis and L1 regularisation. Principal component analysis allowed us to identify 15 principal components that have the greatest impact on classification accuracy and, accordingly, use them in the classification process. Analysing the results, we found that the best classifier was Random Forest with 95.9 % accuracy, and the worst was Naive Bayes with 85.1 %. To evaluate the quality of classification and select the best classifier, the Confusion matrix is used, which takes into account the number of true positive (TP) and true negative (TN) values, as well as the number of false negative (FN) and false positive (FP) classification results, and the values of such metrics as accuracy, precision, sensitivity, F1, and ROC. Accuracy is the percentage of correct answers given by the algorithm, while Recall is the number of TPs divided by the number of TPs plus the number of FNs. F1 indicates the balance between accuracy and sensitivity. Precision is the number of true positive predictions divided by the number of false positive and true negative predictions. ROC AUC is a tool for measuring performance for classification tasks at different thresholds. It shows how well a model can distinguish between classes. The conclusions present the main results of the study and indicate the main future direction of the work, namely, the study of classification results for other datasets and more efficient processing and analysis.

BackgroundProstate cancer screening is controversial because of uncertainty about its benefits and risks. The aim of this survey was to reveal preferences of men concerning prostate cancer screening and to test the effect of an informative video on these preferences.MethodsA stated preferences questionnaire was sent by e-mail to men aged 50–75 with no history of prostate cancer. Half of them were randomly assigned to view an informative video. A discrete choice model was established to reveal men’s preferences for six prostate cancer screening characteristics: mortality by prostate cancer, number of false positive and false negative results, number of overdiagnosis, out-of-pocket costs and recommended frequency.ResultsA population-based sample composed by 1024 men filled in the entire questionnaire. Each attribute gave the expected sign except for overdiagnosis. The video seemed to increase the intention to abstain from prostate cancer screening.ConclusionsThe participants attached greater importance to a decrease in the number of false negatives and a reduction in prostate cancer mortality than to other risks such as the number of false positives and overdiagnosis. Further research is needed to help men make an informed choice regarding screening.

This paper discusses a novel hybrid approach to text classification that integrates a machine learning algorithm along with DistilBERT, a pre-trained deep learning framework for natural language processing, offers a base model fine-tuned on Indian Banking News-Events with a rule-based method that is used by filtering false positives and dealing with false negatives to enhance the results given by the previous classifier. The major benefit is that by incorporating unique rules for certain chaotic or overlapping categories that have not been effectively trained, the system can be quickly fine-tuned. This research also compares the effectiveness of state-of-art deep contextual language representation DistilBERT used in our proposed hybrid model with the most preferred context independent language representation TFIDF, on supervised learning of the classification of multiclass Banking News-Events. Both representations are fed into the machine learning classifiers, Logistic Regression, Linear SVC, Decision Tree, and Random Forest. The findings indicate that DistilBERT can better transfer generic domain knowledge to other domains as compared to the baseline TFIDF and results into higher accuracy with Random Forest. The result is further fed into the Rules based method as per our proposed model. And, our proposed hybrid model resulted into improved accuracy.

Numerical estimation has been used to study how children mentally represent numbers for many years (e.g., Siegler & Opfer, 2003). However, these studies have always presented children with positive numbers and positive number lines. Children’s mental representation of negative numbers has never been addressed. The present study tested children in the 2nd, 4th, and 6th grades to assess their mental representations of both positive and negative numbers using a standard numerical estimation task. We replicated the shift from a logarithmic to linear representation for positive numbers (0–1,000 scale) in that 2nd graders represented positive numbers logarithmically, but 4th and 6th graders represented the numbers linearly. Furthermore, children’s representation of negative numbers paralleled their representations of positive numbers and showed the same shift from a logarithmic representation at Grade 2 to linear representations at Grades 4 and 6. This is the first study to provide data on children’s representation of negative numbers, and the implications of these findings are discussed.

Head movements can greatly affect swallowing accelerometry signals. In this paper, we implement a spline-based approach to remove low frequency components associated with these motions. Our approach was tested using both synthetic and real data. Synthetic signals were used to perform a comparative analysis of the spline-based approach with other similar techniques. Real data, obtained data from 408 healthy participants during various swallowing tasks, was used to analyze the processing accuracy with and without the spline-based head motions removal scheme. Specifically, we analyzed the segmentation accuracy and the effects of the scheme on statistical properties of these signals, as measured by the scaling analysis. The results of the numerical analysis showed that the spline-based technique achieves a superior performance in comparison to other existing techniques. Additionally, when applied to real data, we improved the accuracy of the segmentation process by achieving a 27% drop in the number of false negatives and a 30% drop in the number of false positives. Furthermore, the anthropometric trends in the statistical properties of these signals remained unaltered as shown by the scaling analysis, but the strength of statistical persistence was significantly reduced. These results clearly indicate that any future medical devices based on swallowing accelerometry signals should remove head motions from these signals in order to increase segmentation accuracy.

We sought to assess the accuracy of different search terms to identify individuals with active seasonal allergic rhinitis (SAR) in general practice. A reference search strategy was developed to identify patients with active SAR. This was applied through inspection of electronic health records of patients aged 15-45 years in a 10% random sample of a general practice database. Searches used Read codes and medication relating to SAR. Sensitivity, specificity, and positive and negative predictive values were calculated. Using the reference search strategy, 54/1092 (4.9%) of 15-45 year-old patients had current SAR. Searching for drugs used in nasal allergy had the highest sensitivity (85%) and good specificity (86%). Searching for a recorded history of SAR (H170) in the last two years was more specific (100%) but this approach only had limited sensitivity (17%). Electronic searches can be used to identify patients with current SAR, but the accuracy varies widely. Larger numbers of sufferers can be identified using broader search parameters, but with increasing numbers of false positives. In contrast, more focused search strategies give a smaller yield needing less cleaning of data to identify true positives, but there is an associated increase in the number of false negatives.

Software defects can cause much loss. Static bug-finding tools are believed to help detect and remove defects. These tools are designed to find programming errors; but, do they in fact help prevent actual defects that occur in the field and reported by users? If these tools had been used, would they have detected these field defects, and generated warnings that would direct programmers to fix them? To answer these questions, we perform an empirical study that investigates the effectiveness of state-of-the-art static bug finding tools on hundreds of reported and fixed defects extracted from three open source programs: Lucene, Rhino, and AspectJ. Our study addresses the question: To what extent could field defects be found and detected by state-of-the-art static bug-finding tools? Different from past studies that are concerned with the numbers of false positives produced by such tools, we address an orthogonal issue on the numbers of false negatives. We find that although many field defects could be detected by static bug finding tools, a substantial proportion of defects could not be flagged. We also analyze the types of tool warnings that are more effective in finding field defects and characterize the types of missed defects.

4 - POSITIVES AND NEGATIVES

Despite careful patient selection, induction chemotherapy for acute myeloid leukemia (AML) is associated with a considerable risk for treatment-related mortality (5%-20%). We evaluated machine learning (ML) algorithms trained using factors available at the time of admission for AML therapy to predict death during the hospitalization. We included AML discharges with age > 17 years who received inpatient chemotherapy from State Inpatient Database from Arizona, Florida, New York, Maryland, Washington, and New Jersey for years 2008-2014. The primary objective was to predict inpatient mortality in patients undergoing chemotherapy using covariates present before initiation of chemotherapy. ML algorithms logistic regression (LR), decision tree, and random forest were compared. 29,613 hospitalizations for patients with AML were included in the analysis each with 4,177 features. The median age was 58.9 (18-101) years, 13,689 (53.7%) were male, and 20,203 (69%) were White. The mean time from admission to chemotherapy was 3 days (95% CI, 2.9 to 3.1), and 2,682 (9.1%) died during the hospitalization. Both LR and random forest models achieved an area under the curve (AUC) score of 0.78, whereas decision tree achieved an AUC of 0.70. The baseline LR model with age yielded an AUC of 0.62. To clinically balance and minimize false positives, we selected a decision threshold of 0.7 and at this threshold, 51 of our test set of 5,923 could have potentially averted treatment-related mortality. Using readily accessible variables, inpatient mortality of patients on track for chemotherapy to treat AML can be predicted through ML algorithms. The model also predicted inpatient mortality when tested on different data representations and paves the way for future research.

We re-addressed the question of whether routine total urinary protein determinations can be used to predict the presence of microalbuminuria by studying 61 patients who attended a diabetic clinic and tested negative or had one positive protein by dipstick. Total urinary protein was measured by the Biorad dye-binding method in undialyzed urine (UND), in dialyzed urine (DIAL), and in dialyzed urine in which albumin and globulins were separated, measured separately with albumin and globulin standards and the results added together to obtain total urinary protein (A + G). The results were compared with albumin measurements obtained by radioimmunoassay (RIA). Compared to DIAL, urinary protein measurements were 43% higher with A + G and 22% higher with UND. Microalbuminuria correlated moderately with UND (r = 0.81) and better with the other methods (r = 0.87 for DIAL, r = 0.91 for A + G). None of the methods predicted microalbuminuria reliably. Taking a protein-to-creatinine ratio of 0.15 and an albumin-to-creatinine ratio of 0.03 as upper limits of normal, we found that UND had a 72% positive predictive value (28% false positives) and 85% negative predictive value (15% false negatives). DIAL had 90% positive predictive value (10% false positives) and 78% negative predictive value (22% false negatives). A + G had 65% positive predictive value (35% false positives) but 91% negative predictive value (9% false negatives). A + G, which uses the correct standards, would be the most suitable method for screening, having the least number of false negatives, but has more false positives because it is more sensitive. In practice, most routine chemical laboratories find it expedient to use only UND, but physicians interpreting the results of this method should be aware of its limitations.

Sense the moment: A highly sensitive antimicrobial activity predictor based on hydrophobic moment

Epidemics, such as Dengue and Influenza, remain significant threats to public health, particularly in densely populated areas. These diseases can spread rapidly, posing a challenge to early detection and containment. Existing methods for epidemic prediction often rely on basic surveillance and historical data, but these approaches have limitations, including a lack of real-time updates and the ability to predict disease trends with high accuracy. This project aims to leverage artificial intelligence (AI) and machine learning techniques to predict the infection status of individuals based on blood sample data and visualize epidemic spread patterns. Using medical datasets containing blood parameters such as WBC count, CRP levels, platelet count, and age, the AI model can classify individuals as infected, cured, or deceased, and determine whether the cause is Dengue or Influenza. Machine learning algorithms like Logistic Regression, AdaBoost, ANN, Random Forest, LGBM, and Decision Trees have been applied to the data, resulting in high prediction accuracy. The system also integrates real-time visualizations, showing infection statistics, recovery trends, and mortality rates, and it utilizes population data to predict the regional spread of the epidemic. This project represents a significant advancement over traditional methods by providing a robust, AI-based solution for epidemic prediction and management.

Parkinson’s disease (PD) is a nervous system disorder first described as a neurological condition in 1817. It is one of the more prevalent diseases in the elderly, and Alzheimer’s is the second most common neurodegenerative illness. It impacts the patient’s movement. Symptoms start gradually with tremors, stiffness in movement, and speech and voice disorders. Researches proved that 89% of patients with Parkinson’s has speech disorder including uncertain articulation, hoarse and breathy voice and monotone pitch. The cause behind this voice change is the reduction of dopamine due to damage of neurons in the substantia nigra responsible for dopamine production. In this work, Parkinson’s disease is classified with the help of human voice signals. Six different machine learning (ML) algorithms are used in the classification: Stochastic Gradient Descent (SGD) Classifier, Extreme Gradient Boosting (XGB) Classifier, Logistic Regression Classifier, Random Forest Classifier, K-Nearest Neighbour (KNN) Classifier, and Decision Tree (DT) Classifier. This research aims to classify Parkinson’s disease using human voice signals and extract essential features to reduce the complexity of the dataset. Then, human voice signals are analyzed to check the voice intensity and spectrum for PD patients. Then, machine learning classifiers are applied to classify the PD patients based on the extracted features. The results show that SGD-Classifier has 91% accuracy, XGB-Classifier has 95% accuracy, Logistic Regression has 91% accuracy, Random Forest shows 97% accuracy, KNN shows 95% accuracy, and Decision Tree has 95% accuracy. Hence, Random Forest has the highest accuracy. The disease can be studied more by looking for more characteristics of PD patients to enhance its proper use in the medical field.