Data gold mining

The Motorway Incident Detection and Automatic Signalling (MIDAS) system data are summaries of traffic volumes, speeds and vehicle types, derived from loop detectors that sense traffic passing over them. These sensors are being rolled out across the UK's strategic highway network (motorways and major A-roads) and data from portions of one of the busiest motorways, the M25, now go back 10 years. Richard Gibbens and Wiebke Werft describe a journey time predictor based on MIDAS and explain how the data can shed light on topical issues such as road pricing.     

数据挖掘及其对统计学的挑战

一、引言随着科学技术的发展 ,利用数据库技术来存储管理数据 ,利用机器学习的方法来分析数据 ,从而挖掘出大量的隐藏在数据背后的知识 ,这种思想的结合形成了现在深受人们关注的非常热门的研究领域 :数据库中的知识发现———ＫＤＤ(ＫｎｏｗｌｅｄｇｅＤｉｓｃｏｖ ｅｒｙｉｎＤａｔａｂａｓｅｓ) ,其中 ,数据挖掘技术便是ＫＤＤ中的一个最为关键的环节。1995年 ,在加拿大的蒙特利尔召开了第一届“知识发现和数据挖掘”国际学术会议 ,数据挖掘一词被很快流传开来。数据挖掘———ＤＭ(ＤａｔａＭｉｎｉｎｇ)就是从大量的、不完全的、有噪…     

The Design of and R Control Charts for Skew Normal Distributed Data

The Shewhart-type control chart is traditionally developed under the normality assumption. In practice, however, this assumption may not hold. Because the skew normal distribution represents a broad distribution class and is more flexible than is the normal distribution, we propose two new control charts to monitor process mean and spread for skew normal distributed data. Moreover, to facilitate practical implementation, tables of charting constants are provided. We conducted simulation studies to compare the false alarm rates, and the results show that new proposed charts perform better than others as skewness increases. Finally, an illustrative example is provided.     

MultiOrd: An R Package for Generating Correlated Ordinal Data

In this article, operational details of an R package MultiOrd that is designed for the generation of correlated ordinal data are described, and examples of some important functions are given. The package provides a valuable and needed tool that has been lacking for generating multivariate ordinal data.     

BinNor: An R Package for Concurrent Generation of Binary and Normal Data

This article describes the package BinNor, which is designed for generating multiple binary and normal variables simultaneously given marginal characteristics and association structure via combining well-established results from the random number generation literature, based on the methodology proposed by Demirtas and Doganay.