ADAPTIVE TRIMMED MEAN AS A LOCATION ESTIMATE

The trimmed mean is one of the most common estimators of location for symmetrical distributions, whose effect depends on whether the trim rate matches the proportion of contaminated data. Based on the geometric characteristics of the curve of the trimmed variance function, the authors propose two kinds of adaptive trimmed mean algorithms. The accuracy of the estimators is compared with that of other often-used estimates, such as sample mean, trimmed mean, trimean, and median, by means of simulation method. The results show that the accuracy of the adaptive derivative optimization trimmed mean method is close to the optimum performance in case of medium contamination （the contamination rate is less than 50%）. Under high contamination situation （the contamination rate equals 8070）, the performance of the estimates is comparable to that of the median and is superior to other counterparts.

Enhanced Detection of Cerebral Atherosclerosis Using Hybrid Algorithm of Image Segmentation

In medical science for envisaging human body’s phenomenal structure a major part has been driven by image processing techniques.Major objective of this work is to detect of cerebral atherosclerosis for image segmentation applica-tion.Detection of some abnormal structures in human body has become a difficult task to complete with some simple images.For expounding and distinguishing neural architecture of human brain in an effective manner,MRI(Magnetic Reso-nance Imaging)is one of the most suitable and significant technique.Here we work on detection of Cerebral Atherosclerosis from MRI images of patients.Cer-ebral Atherosclerosis is a cerebral vascular disease causes narrowing of the arteries due to buildup of fatty plaque inside the blood vessels of the brain.It leads to Ischemic stroke if not diagnosed early.Stroke affects majorly old age people and percentage of affected women is more compared to men.Results:Preproces-sing is done by using alpha trimmed meanfilter which is used to remove noise and also it enhances the image.Segmentation of cerebral atherosclerosis is done by using K-means clustering,Contextual clustering,and proposed Hybrid algo-rithm.Various parameters like Correlation,Pixel density,energy is determined and from the analysis of parameters it is determined that proposed Hybrid algo-rithm is efficient.

Analysis of Salaries and Some Non-traditional Measures of Location

The paper deals with an analysis of how to use certain measures of location in analysis of salaries. One of the traditional measures of location, the mean should offer typical value of variable, representing all its values by the best way. Sometimes, the mean is located in the tail of the distribution and gives a very biased idea about the location of the distribution. In these cases, using different measures of location could be useful. Trimmed mean is described. The trimmed mean refers to a situation where a certain proportion of the largest and smallest observations are removed and the remaining observations are averaged. The construction of some measures of location is based on the analysis of outliers. Outliers are characterized. Then the possibilities of the detection of outliers are analyzed. Computing of one-step M-estimator and modified one-step M-estimator of location is described. A comparison of the trimmed means and M-estimators of location is presented. Finally, the paper focuses on the application of the trimmed mean and M-estimators of location in analysis of salaries. The analysis of salaries of employers of the big Slovak companies in second half of the year 2009 is realized. The data from the census are used in the analysis. The median, 20% trimmed mean and the characteristics, based on the one-step M-estimator of location and modified one step M-estimator, are calculated.

Trimmed and Winsorized Transformed Means Based on a Scaled Deviation

This paper introduces the Tukey trimmed and Winsorized means for the transformed data based on a scaled deviation. The trimmed and Winsorized means and scale based on a scaled deviation are as special cases. Meanwhile, the trimmed and Winsorized skewness and kurtosis based on a scaled deviation are given.Furthermore, some of their robust properties（influence function, breakdown points） and asymptotic properties（asymptotic representation and limiting distribution） are also obtained.

Choice of Optimal Trimming Proportion by the Random Weighting Method

In this paper, a strongly consistent estimation of the optimal trimming proportion in trimmed mean is found by the random weighting method. In addition, using the same method a strongly consistent estimation for the distribution of some adaptive estimator is also obtained.