lecture 1

1. not double pass, all homeworks submitted through give. All exams questions are short answers

2. consultation is Friday through zoom from 1 to 2

3. major characteristics of big data: volume, variety, velocity, value, visibility, variability, veracity

数量(size),种类,速度,值,可见性,可变性,准确性

4. volume

quantity of data being created from all sources

带来的问题是存储空间不够以及time complexity

volume增加cost增加

5. variety

1) different types:

relational data( tables ransactions ) with structures, fixed schema

text data( books, reports) unstructures

semi-structured data(JSON, XML ) has orgnizations as well

graph data( social network, RDF )

imagevideo data( Instagram, Youtube )

2) different sources

实际应用中,一个app会有多种来源或由不同种类的信息组合成的

data integration的主要问题为Heterogeneous,data integration将来源不用的信息组合成一个独特的view,Heterogeneous是指the schema of view is different with each other。传统的解决方法是schema mapping。解决难度和时间复杂度与the level of heterogentity和data sources有关。另一个问题是record linkage in variety data,指wether 2 records refer to the same entity or not,需要我们尽可能详尽的使用来源不同的各种不同信息

data curation指organization and integration of data collected from various sources,可能出现的问题是long tail of data variety

data curation即使数据更有序可以减少long tail

6. velocity( speed )

很多应用需要及时回馈

需要解决的是batch processing, real time processing与transmission

原文地址:https://www.cnblogs.com/eleni/p/13032018.html