THE OUTLIER TEST BASED ON NORMALLY DISTRIBUTED DATA
Journal Title: Topics in Intelligent Computing and Industry Design (ICID) - Year 2017, Vol 1, Issue 3
Abstract
In this paper, the simulation experiments are used to research the efficiency of different outlier test. R software is used to generate the normally distributed data, which contains the outliers defined differently. Then each model are applied to test the outliers in the dataset. The experimental result shows that the efficiencies of the same model differed greatly when the definition of the abnormal value varies. So we can draw the conclusion that the efficiency of the test model depends on the definition of the outlier. And the efficiency of Quartile model is higher than other models. So Quartile model is more effective and applicable than other models in actual data test.
Authors and Affiliations
Chen Peifan
FIELD MEASUREMENT AND ANALYSIS OF VIBRATION REDUCTION EFFECT OF STEEL SPRING FLOATING PLATE AND VIBRATION RESPONSE STUDY ON SURROUNDING BUILDING
In order to realize the vibration reduction effect of steel-spring floating slab to surrounding buildings, by taking the rail transit underground of a city as the research object, three sites, with the ordinary track bed...
RESEARCH ON BUFFER OVERFLOW ATTACK EFFECT EVALUATION TECHNOLOGY
Buffer overflow is a very common and serious consequences of security vulnerability. Attack effect evaluation plays an important role in software security assessment. Proceeding from the two aspects of the buffer overflo...
THE DESIGN OF BABY AUTOMATICALLY WOBBLES THE SLEEPING DEVICE
In twenty-first Century, people’s living demand is higher and higher, the SCM application is more and more widespread, and the residents’ equipment is constantly intelligent, in which the baby automatic rocking sleep dev...
DESIGN OF HIGHLY RELIABLE LOGIC CONTROL SOFTWARE FOR ESCAPE SYSTEM IN PRE-LAUNCHING STAGE
In order to meet the sudden danger of manned launch vehicle on the ground, emergency escape control design in the pre-launching stage should ensure high reliability and high safety. Based on programmable logic control co...
A NOVEL DATA CLASSIFICATION APPROACH USING FUZZY LOGIC
Data Mining is the core part of knowledge discovery aims to extract hidden patterns from data. It is mainly used in information processing applications. It is known that the classification techniques divide data based on...