scholarly journals Binary Classifier Calibration Using an Ensemble of Linear Trend Estimation

2016 ◽  
Author(s):  
Mahdi Pakdaman Naeini ◽  
Gregory F. Cooper
Keyword(s):  
Linear Trend ◽  
Binary Classifier ◽  
Trend Estimation ◽  
Classifier Calibration

scholarly journals Robust Trend Estimation for AR(1) Disturbances

2016 ◽  
Vol 34 (2) ◽  
Author(s):  
Roland Fried ◽  
Ursula Gather
Keyword(s):  
Robust Estimation ◽  
Linear Trend ◽  
Autoregressive Process ◽  
Trend Estimation ◽  
First Order ◽  
Repeated Median

We discuss the robust estimation of a linear trend if the noise follows an autoregressive process of first order. We find the ordinary repeated median to perform well except for negative correlations. In this case it can be improved by a Prais-Winsten transformation using a robust autocorrelation estimator.

scholarly journals Trend Analysis of Temperature and Rainfall across Agro Climatic Zones of Karnataka-A Semi Arid State in India

2021 ◽  
pp. 67-78
Author(s):  
Seedari Ujwala Rani ◽  
Naveen P. Singh ◽  
Pramod Kumar ◽  
Rabindra Nath Padaria ◽  
Ranjit Kumar Paul
Keyword(s):  
Linear Trend ◽  
Rainfall Anomaly ◽  
Maximum Temperature ◽  
Climatic Zones ◽  
Trend Estimation ◽  
North Eastern ◽  
Karnataka State ◽  
Dry Zone ◽  
Rainfall Anomalies ◽  
Semi Arid

The study was carried out for ten Agro climatic zones in Karnataka state in India. The temperature and rainfall data were used for analysis from 1979-2019 which is about 40 years. Understanding spatiotemporal rainfall pattern, Rainfall Anomaly Index which is drought indicator technique was  used to classify the positive and negative severities in rainfall anomalies. The RAI ranges below 0.2 are considered as dry zone. The analysis resulted that, all zones are falls in category of dry zone with range of 0.2 to 0.4. For past five years, North Eastern Transition Zone was noted maximum times falling in the range of RAI below 0.2 and near to zero. Statistical techniques like linear trend estimation, R square was used for trend estimation across annual, seasonal to identify the variation in the temperature across different zones. The meaningful statistically significant achieves when there is r2≥0.65 and p≤0.05. It was analysed that, hilly Zone experienced decreased trend in both minimum and maximum temperature in all seasons which ultimately reflected in annual temperature to decrease with high R square values.

ASSESSMENT OF STATISTICAL SIGNIFICANCE OF HISTORIC DANUBE FLOODS

2020 ◽  
Author(s):  
Stevan Prohaska ◽  
◽  
Aleksandra Ilić ◽  
Pavla Pekarova ◽  
◽  
...  
Keyword(s):  
Middle Ages ◽  
Linear Trend ◽  
Statistical Significance ◽  
Danube River ◽  
Exceedance Probability ◽  
Simultaneous Occurrence ◽  
Flood Wave ◽  
Probability Of Occurrence ◽  
Comprehensive Method ◽  
Flood Hydrograph

Data on historic floods along the Danube River exist since the year 1012. In the Middle Ages, floods were estimated based on historical documents, including original handwritten notes, newspaper articles, chronicles, formal letters, books, maps and photographs. From 1500 until the beginning of organized water regime observations, floods were hydraulically reconstructed based on water marks on old buildings in cities along the Danube (Passau, Melk, Emmersdorf an der Donau, Spilz, Schonbuhen and Bratislava). The paper presents a procedure for assessing the statistical significance of registered historic floods using a comprehensive method for defining theoretical flood hydrographs at hydrological stations. The approach is based on correlation analysis of two basic flood hydrograph parameters – maximum hydrograph ordinate (peak) and flood wave volume. The PROIL model is used to define the probability of simultaneous occurrence of these parameters. It defines the exceedance probability of two random variables, in the specific case two hydrograph parameters of the form: P{Qmax more equal to qmax,p)∩(Wmax more equal to wmax,p)} = P (1) where: Qmax – maximum hydrograph ordinate (peak); qmax,p – maximum discharge of the probability of occurrence p; Wmax – maximum hydrograph volume; wmax,p – maximum flood wave volume of the probability of occurrence p; P – exceedance probability. Spatial positions of the lines of exceedance of two flood hydrograph parameters and the empirical points of the corresponding parameters of the considered historic flood in the correlation field Qmax - Wmax, allow direct assessment of the exceedance probability of a historic flood, or its statistical significance. The proposed procedure was applied in practice to assess the statistical significance of the biggest floods registered along the Danube in the sector from its mouth to the Djerdap 1 Dam. The linear trend in the time-series of maximum annual flows at a representative hydrological station and the frequency of historic floods in the considered sector of the Danube are discussed at the end of the paper.

Identifying changes in trends in the age standardized incidence of melanoma in Australia

2019 ◽  
Vol 3 (4) ◽  
pp. 250-252 ◽  
Author(s):  
David M Hille
Keyword(s):  
Public Health ◽  
Linear Regression ◽  
Iterative Process ◽  
Linear Trend ◽  
Piecewise Linear ◽  
Health Issue ◽  
Public Health Issue ◽  
Piecewise Linear Regression ◽  
Significant Public Health ◽  
Males And Females

ObjectiveTo identify changes in the linear trend of the age-standardized incidence of melanoma in Australia for all persons, males, and females. MethodsA two-piece piecewise linear regression was fitted to the data. The piecewise breakpoint varied through an iterative process to determine the model that best fits the data.ResultsStatistically significant changes in the trendof the age-standardized incidence of melanoma in Australia were found for all persons, males, and females. The optimal breakpoint for all persons and males was at 1998. For females, the optimal breakpoint was at 2005. The trend after these breakpoints was flatter than prior to the breakpoints, but still positive.ConclusionMelanoma is a significant public health issue in Australia. Overall incidence continues to increase. However, the rate at which the incidence is increasing appears to be decreasing.

DAMPAK VOLATILITAS HARGA MINYAK DI INDONESIA TERHADAP PENYESUAIAN KESEIMBANGAN FISKAL

2015 ◽  
Vol 15 (2) ◽  
pp. 37-52
Author(s):  
Mahpud Sujai
Keyword(s):  
Linear Trend ◽  
The State ◽  
Oil Price ◽  
Output Gap ◽  
State Budget ◽  
Price Changes ◽  
Potential Output ◽  
Budget Cycle ◽  
Actual Output ◽  
The Impact

This paper is intended to analyze the effect of oil price changes on potential output and actual output in the state budget cycle and identifies the output gap which is the difference between potential output and actual output. The research methodology uses a quantitative approach to analyze problems that occur related to the impact of oil price changes to the state budget cycle. Data analysis was carried out through the approach cyclically adjusted fiscal balance with a simplified approach. This research identified that the potential output is likely to continue increasing in line with Indonesia's oil price trends which is continue to rise following the world oil price movements. In calculating the output gap using a linear trend and HP filter, the result is fuctuating depend on the percentage changes in both potential output and actual output. This paper concludes that Indonesian oil price (ICP) has a significant impact on changes in the state budget cycle. If oil prices rise, the output gap between potential output and actual output is greater, and vice versa. This will make the budget vulnerable to shock that occurs as an external infuence.

ANALYSIS OF THE STRUCTURE AND PREDICTION OF HIV INFECTION INCIDENCE IN THE VORONEZH REGION

2020 ◽  
pp. 147-152
Author(s):  
Евгений Николаевич Коровин ◽  
Юлия Викторовна Сиромашенко ◽  
Владимир Николаевич Коровин
Keyword(s):  
Hiv Infection ◽  
Graphical Representation ◽  
Linear Trend ◽  
Risk Groups ◽  
Extrapolation Methods ◽  
The Past ◽  
Voronezh Region ◽  
The Future ◽  
Dynamic Series ◽  
Exponential Smoothing Method

В статье приведены анализ и прогнозирование основных статистических показателей, характеризующих развитие эпидемической ситуации по вирусу иммунодефицита человека (ВИЧ) в Воронежской области, а именно распределение по путям инфицирования. В качестве данных для прогнозирования были использованы показатели заболеваемости прошлых лет. Применяемый метод прогнозирования основан на методах экстраполяции. Все методы экстраполяции объединяет то, что они проецируют на будущее ход событий, сложившийся в прошлом. При этом не устанавливаются никакие причинные связи - принимается, что действующие в прошлом силы без существенных изменений будут действовать и в будущем. При формировании прогнозов с помощью экстраполяции исходят из статистически складывающихся тенденций изменения тех или иных количественных характеристик объекта за определённый период. Прогнозирование заболеваемости ВИЧ осуществляется с помощью метода экспоненциального сглаживания с использованием линейного тренда и выбором оптимальных параметров сглаживания. Трендом называют аналитическое или графическое представление изменения переменной во времени, полученное в результате выделения регулярной (систематической) составляющей динамического ряда. Основной целью анализа и прогнозирования является выявление основных тенденций по распространению ВИЧ-инфекции, определение основных путей инфицирование, а также выделение ключевых групп риска среди населения Воронежской области The article analyzes and predicts the main statistical indicators that characterize the development of the epidemic situation of human immunodeficiency virus (HIV) in the Voronezh region, namely, the distribution by infection pathways. Previous years ' morbidity rates were used as data for forecasting. The applied forecasting method is based on extrapolation methods. What all extrapolation methods have in common is that they project the course of events in the past into the future. At the same time, no causal relationships are established - it is assumed that the forces operating in the past will continue to operate in the future without significant changes. When forming forecasts using extrapolation, they are based on statistically developing trends in changes in certain quantitative characteristics of an object over a certain period. Predicting the incidence of HIV is carried out using the exponential smoothing method using a linear trend and choosing the optimal smoothing parameters. A trend is an analytical or graphical representation of changes in a variable over time, resulting from the allocation of a regular (systematic) component of a dynamic series. The main goal of the analysis and forecasting is to identify the main trends in the spread of HIV infection, identify the main routes of infection, and identify key risk groups among the population of the Voronezh region

Sign in / Sign up

Export Citation Format

Share Document