Predicting graft survival in paediatric kidney transplant recipients using machine learning.

BACKGROUND: Identification of factors that affect graft survival in kidney transplantation can increase graft survival and reduce mortality. Artificial intelligence modelling enables impartial evaluation of clinician bias. This study aimed to examine factors that affect the survival of grafts in paediatric kidney transplantation through the use of machine learning. METHODS: A retrospective review was conducted on records of paediatric patients who underwent kidney transplantation between 1994 and 2021 and had post-transplant follow-up > 12 months. The nearest neighbour method was used to impute missing fields from a total of 48 variables in the dataset. Models including Naive Bayes, logistic regression, support vector machine (SVM), multi-layer perceptron, and XGBoost were trained to predict graft survival. The study used 80% of the patients for training and the remaining 20% for testing. Modelling success was evaluated based on accuracy and F1 score metrics. RESULTS: The study analysed 465 kidney transplant recipients. Of these, 56.7% were male. The mean age at transplantation was 12.08 ± 5.01 years. Of the kidney transplants, 73.1% (n = 339) were from living donors, 34.5% (n = 160) were pre-emptive transplants, and 2.2% (n = 10) were second-time transplants. The machine learning model identified several features associated with graft survival, including antibody-mediated rejection (+ 0.7), acute cellular rejection (+ 0.66), eGFR at 3 years (+ 0.43), eGFR at 5 years (+ 0.34), pre-transplant peritoneal dialysis (+ 0.2), and cadaveric donor (+ 0.2). The successes of the logistic regression and SVM models were similar. The F1 score was 91.9%, and accuracy was 96.5%. CONCLUSION: Machine learning can be used to identify factors that affect graft survival in kidney transplant recipients. By expanding similar studies, risk maps can be created prior to transplantation.

Automated Bird Counting with Deep Learning for Regional Bird Distribution Mapping.

A challenging problem in the field of avian ecology is deriving information on bird population movement trends. This necessitates the regular counting of birds which is usually not an easily-achievable task. A promising attempt towards solving the bird counting problem in a more consistent and fast way is to predict the number of birds in different regions from their photos. For this purpose, we exploit the ability of computers to learn from past data through deep learning which has been a leading sub-field of AI for image understanding. Our data source is a collection of on-ground photos taken during our long run of birding activity. We employ several state-of-the-art generic object-detection algorithms to learn to detect birds, each being a member of one of the 38 identified species, in natural scenes. The experiments revealed that computer-aided counting outperformed the manual counting with respect to both accuracy and time. As a real-world application of image-based bird counting, we prepared the spatial bird order distribution and species diversity maps of Turkey by utilizing the geographic information system (GIS) technology. Our results suggested that deep learning can assist humans in bird monitoring activities and increase citizen scientists' participation in large-scale bird surveys.