Search

Estimating class proportions has emerged as an important direction in positive-unlabeled learning. Well-estimated class priors are key to accurate approximation of posterior distributions and are necessary for the recovery of true classification performance. While significant progress has been made in the past decade, there remains a need for accurate strategies that scale to big data. Motivated by this need, we propose an intuitive and fast nonparametric algorithm to estimate class proportions. Unlike any of the previous methods, our algorithm uses a sampling strategy to repeatedly (1) draw an example from the set of positives, (2) record the minimum distance to any of the unlabeled examples, and (3) remove the nearest unlabeled example. We show that the point of sharp increase in the recorded distances corresponds to the desired proportion of positives in the unlabeled set and train a deep neural network to identify that point. Our distance-based algorithm is evaluated on forty datasets and compared to all currently available methods. We provide evidence that this new approach results in the most accurate performance and can be readily used on large datasets.

Positive-unlabeled learning is often studied under the assumption that the labeled positive sample is drawn randomly from the true distribution of positives. In many application domains, however, certain regions in the support of the positive class-conditional distribution are over-represented while others are under-represented in the positive sample. Although this introduces problems in all aspects of positive-unlabeled learning, we begin to address this challenge by focusing on the estimation of class priors, quantities central to the estimation of posterior probabilities and the recovery of true classification performance. We start by making a set of assumptions to model the sampling bias. We then extend the identifiability theory of class priors from the unbiased to the biased setting. Finally, we derive an algorithm for estimating the class priors that relies on clustering to decompose the original problem into subproblems of unbiased positive-unlabeled learning. Our empirical investigation suggests feasibility of the correction strategy and overall good performance.

BACKGROUND: The Nottingham Clavicle Score (NCS) has been recently described for functional outcome assessment after injuries to clavicle and the Acromioclavicular joint. However; validity and responsiveness are context specific psychometric terms and the NCS has not been previously described in surgically treated clavicle shaft fractures. Aim of the present study was to investigate validity and responsiveness of the NCS in clavicle fractures treated with titanium flexible nailing. METHODS: This prospective study was undertaken on consecutively operated clavicle shaft fractures treated with titanium elastic nail from November 2013 to August 2016. Functional assessment using NCS was done at two and six months postoperatively. Construct validity was also evaluated by formulating the null hypothesis that there would be no difference in NCS at six months after open and closed reduction and in 15B1 and 15B2 fracture sub-types. The above two hypotheses were formulated based on previous studies that used Constant score and DASH score. Pre-specified hypothesis and results in accordance with the hypotheses suggest satisfactory construct validity. Responsiveness was evaluated using standardized response mean (SRM) and Effect size (ES). ES and SRM values ≥0.80 suggest satisfactory responsiveness. The proportion of patients having the least possible score of 0 points (floor effect) and the highest possible score of 100 points (ceiling effect) was evaluated at two and six months postoperatively. Floor and ceiling effect of

Accurately estimating performance accuracy of machine learning classifiers is of fundamental importance in biomedical research with potentially societal consequences upon the deployment of best-performing tools in everyday life. Although classification has been extensively studied over the past decades, there remain understudied problems when the training data violate the main statistical assumptions relied upon for accurate learning and model characterization. This particularly holds true in the open world setting where observations of a phenomenon generally guarantee its presence but the absence of such evidence cannot be interpreted as the evidence of its absence. Learning from such data is often referred to as positive-unlabeled learning, a form of semi-supervised learning where all labeled data belong to one (say, positive) class. To improve the best practices in the field, we here study the quality of estimated performance in positive-unlabeled learning in the biomedical domain. We provide evidence that such estimates can be wildly inaccurate, depending on the fraction of positive examples in the unlabeled data and the fraction of negative examples mislabeled as positives in the labeled data. We then present correction methods for four such measures and demonstrate that the knowledge or accurate estimates of class priors in the unlabeled data and noise in the labeled data are sufficient for the recovery of true classification performance. We provide theoretical support as well as empirical evidence for the efficacy of the new performance estimation methods.

Background Pulpotomy is the treatment for cariously exposed vital primary molars. UsingfFormocresol as a pulpotomy agent is been in controversy, which has triggered the search for better alternatives. The product like 'Myristica fragrans (MF)-Nutmeg gel, Terminalia chebula (TC)-Myrobolan gel is gaining popularity as herbal pulpotomy agents. Aim To evaluate and compare clinical and radiographical success of herbal gels Myristica fragrans (MF)-Nutmeg, and Terminalia chebula (TC)-Myrobolan as pulpotomy medicaments in primary teeth. Materials and methods Twenty participants (n = 20), each with at least two primary molars requiring pulpotomy, were selected and divided into two test groups. In 10 children Terminalia chebula gel was placed in one side and Formocresol on another side. Rest 10 children were treated with Myristica fragrans gel on one side and another side with formocresol. The treated teeth selected for clinical and radiographic evaluation were monitored periodically for 3, 6 and 12 months. Results With the follow-up of 12 months there was no significant difference in efficacy of all three pulpotomy medicaments, i.e. Nutmeg, Myrobolan, and Formocresol, respectively was found. Conclusion Herbal gels have a promising role in dentistry having the proper knowledge, and their effects on teeth would prove them as a successful dental therapeutic agent. How to cite this article Mali S, Singla S, Sharma A, Gautam A, Niranjan B, Jain S. Efficacy of Myristica fragrans and Terminalia chebula as Pulpotomy Agents in Primary Teeth: A Clinical Study. Int J Clin Pediatr Dent, 2018;11(6):505-509.

Background: Tuberculosis is an infectious disease cause by the mycobacterium tuberculosis which typically involve lung but can affect other site called extrapulmonary tuberculosis (EPTB). In EPTB lymph node tuberculosis is most common and most severe is tuberculous meningitis. Cartridge based nucleic acid amplification test (CBNAAT) have high sensitivity and perform both respiratory and non-respiratory specimen. The aim of this retrospective study is to find out the proportion of extra pulmonary tuberculosis in hospitalized patient and determine the association of CBNAAT in diagnosis EPTB and Rifampicin resistance.Methods: This is the retrospective observational study done at Civil Hospital Ahmedabad in between October 2017 to October 2019 among the admitted patient age 1 month to 12 year diagnosed cases of EPTB according to RNTCP guidelines. Data collection done by medical records as indoor case sheets and investigation data from laboratory department.Results: Incidence of EPTB among the hospitalized children is 1.95% and among total tuberculosis patient is 42.48%. Among 191 suspected EPTB cases 59 confirmed indicate sensitivity of 37.3% and no false positive cases indicate 100% positive predictive value. Sensitivity of CBNAAT was highest in gastric aspirate followed by CSF in present study.Conclusions: Tuberculosis is more common in male child less than 5-year-old with frequent risk factor is rural residential area, lower socioeconomic class and contact with active cases. In my study is CNS tuberculosis is most common. In my study mortality is higher among 1-to-5-year age group with co morbid condition are SAM and septicemia.

Intraosseous ganglion cyst is a rare bone tumor and the lesion could often be missed. The diagnosis could be delayed so proper radiologic investigation and index of suspicion is necessary .Differential diagnoses of painful cystic radiolucent carpal lesion are osteoid osteoma, osteoblastoma and intraosseous ganglion. Curettage of the scaphoid lesion and filling of void with bone graft provides good functional outcomes. The cyst contains mucoid viscous material without epithelial or synovial lining. We present a case of 30 years old male with intraosseous ganglion cyst of scaphoid which was treated with curettage and bone grafting. Rarely ganglion cyst is found in small bones of hand and should be considered as differential diagnosis of chronic radial wrist pain.

Hoffa’s fracture is coronal oriented fracture of distal femur with the fracture line extending through the femoral condyles. Hoffa’s fracture is rarely associated with ipsilateral femur and tibia fractures. Proper clinical examination and radiographic evaluation is necessary to diagnose associated injuries around the knee joint with Hoffa’s fracture. Closed reduction of dislocated patella in emergency room and field triage should be avoided to prevent patellar tendon incarcenation, patellar tendon rupture and osteochondral damage. We report a rare case of Hoffa’s fracture with irreducible patellar dislocation and tibial intercondylar eminence fracture following road traffic accident.