Automated Grading of Diabetic Retinopathy with Ultra-Widefield Fluorescein Angiography and Deep Learning

Purpose. The objective of this study was to establish diagnostic technology to automatically grade the severity of diabetic retinopathy (DR) according to the ischemic index and leakage index with ultra-widefield fluorescein angiography (UWFA) and the Early Treatment Diabetic Retinopathy Study (ETDRS) 7-standard field (7-SF). Methods. This is a cross-sectional study. UWFA samples from 280 diabetic patients and 119 normal patients were used to train and test an artificial intelligence model to differentiate PDR and NPDR based on the ischemic index and leakage index with UWFA. A panel of retinal specialists determined the ground truth for our data set before experimentation. A confusion matrix as a metric was used to measure the precision of our algorithm, and a simple linear regression function was implemented to explore the discrimination of indexes on the DR grades. In addition, the model was tested with simulated 7-SF. Results. The model classification of DR in the original UWFA images achieved 88.50% accuracy and 73.68% accuracy in the simulated 7-SF images. A simple linear regression function demonstrated that there is a significant relationship between the ischemic index and leakage index and the severity of DR. These two thresholds were set to classify the grade of DR, which achieved 76.8% accuracy. Conclusions. The optimization of the cycle generative adversarial network (CycleGAN) and convolutional neural network (CNN) model classifier achieved DR grading based on the ischemic index and leakage index with UWFA and simulated 7-SF and provided accurate inference results. The classification accuracy with UWFA is slightly higher than that of simulated 7-SF.

A STUDY OF GROWTH RATE OF WADARA CALVES IN THE SUB-SAHELIAN REGION OF BORNO STATE

Data on birth weight and growth rate from birth to 12 months of age of 90 Wadara calves maintained at the University of Maiduguri Livestock Farm from 1980 - 1987 period were studied. Mean birth weight was 25.5±2.6kg; 26.7 ±2.04kg for 40 males and 24.3±3.19kg for 50 females. The linear regression function was fitted on monthly body weight in two parts: 0-6 months and 6-12 months. Maximum gain in body weight was attained during 6 to 12 months of age in the male (8.30±0.50kg) and 0 to 6 months in the female (0.54±1.07kg). The difference in the rate of gain between the periods to 6 and 6 to 12 months in each of the sexes was highly significant (P< 0.01). Body weights at birth and at various ages were higher (P < 0.05) for the male than the female calves. The relative growth rate was rapid in both sexes during the first 3 months and was appreciable to the 7th month of age before declining progressively. The value in the first quarter was 18.0± 1.86 and 20.2 ± 1.60 percent for the male and female calves respectively.

METODA ZANURZANIA REGRESJI W PRZYPADKU WYSTĘPOWANIA OBSERWACJI NIETYPOWYCH

This paper presents the application of the regression maximum depth for the estimation of linear regression function structural elements. For two-dimensional sets including untypical observations, regression functions were developed using the classical least squares method and a method based on the concept of observation depth measure in a sample. The effect of untypical observations on the estimated models has been noted.

Data Mining Predictive Modeling for Prediction of Gold Prices Based on Dollar Exchange Rates, Bi Rates and World Crude Oil Prices

Gold is an investment instrument that is quite safe from inflationary attacks, and gold is one aspect of initiating investment. Can by buying gold in physical form and then selling when the price has risen high or by digitally investing gold. One of them is by trading gold online. To maximize the benefits of gold trading, a gold price prediction (XAUUSD) is needed for traders. This study aims to (1) Analyze various factors that influence the price of gold (2) Provide recommendations about the prediction of gold prices. Materials that will be used as objects of research to produce gold price predictions include historical XAUUSD (Gold) data itself, historical crude oil data, historical dollar data (USD IDR) and BI 7-Day Repo Rate (BI Rate). ), in producing the prediction of the gold price used Mining Predictive Modeling data using the linear regression function. The results to be achieved from this study is to provide accurate gold price predictions so that it can be used as a reference in making decisions to buy / sell positions in trading. The prediction of the XAUUSD gold price generated is expected to provide significant interest to the investment players (traders) in order to maximize the profit generated.From the results of the trading tests that have been carried out, the implementation of predictive modeling data mining using a linear regression function produces recommendations for gold price predictions (XAUUSD) with an accuracy of 85%.

LONG -TERM CHANGES OF MAIN INPUT -OUTPUT PRICES IN THE HUNGARIAN BROILER SECTOR

The article presents changes of the main input-output prices in the Hungarian broiler industry over a period of 30 years, and associated correlations. For the processing of long-term data, a linear regression function, correlation and regression analysis were used. The cereal prices correlate and their changes also correspond with a change in compound feed prices. A close correlation can be found between cereal price and broiler price, whilst the correlation shown between the compound feed price and broiler price is very close. During the examined period, the feed prices increased at a higher rate than the broiler price. It was also established that the current feed and energy price significantly affect day-old chick prices which corresponds with an increase in price of the broiler. Furthermore, a close relation can be found between energy and feed compound prices.

Longitudinal Evaluation of Radiopacity of Resin Composites: Influence of Photoactivation and Accelerated Photoaging

SUMMARY This study aimed to assess longitudinally the radiopacity of resin composites under the influence of photoactivation and photoaging processes. Ten specimens (1 mm thick and 4 mm in diameter) of three different microhybrid resin composites, Filtek Z250 XT (R1), TPH 3 Spectrum (R2), and Opallis (R3), were prepared for this study. For longitudinal assessment of radiopacity, radiographic images were obtained five times. The first time (T1), the specimens were not photoactivated; the second time (T2), the specimens were photoactivated; and the next three times, photoaging was carried out, with images obtained at 24 hours (T3), 48 hours (T4), and 72 hours (T5) after this process. The photoaging was conducted using LED light (700 lumens, 9 W, 6400 k) under controlled environmental conditions at 37°C (±1°C) and 65% (±5%) relative humidity. The digital system DIGORA Optime was used. The digital images were measured using the histogram function, and then the pixel intensity values were converted into mmAl (the standard unit of radiopacity) using a linear regression function, with minimal adjustment of R2 ≥ 0.9. Data in mmAl were statistically analyzed using an analysis of variance (α=0.05). R2 resin composite showed higher values of radiopacity, R1 resin composite showed intermediate values, and R3 resin composite showed lower values. Only at T1 did the higher radiopacity of R2 composite differ significantly from other groups (p = 0.0000). After application of treatments (photoactivation and photoaging), all radiopacity values were similar (p-values to T2=0.0507, T3=0.0536, T4=0.0502, T5=0.0501) due to consecutive increase of radiopacity of R1 and R3 composites from T2. Photoactivation and photoaging processes influenced the radiopacity, but changes occurring in the degree of radiopacity were dependent on the composition and chemical characteristics of each composite used.