大数据的类型。

结构化数据非结构化数据以及半结构化数据。
Varieties of big data.
It is divided into three types of big data. The first one is about the structured data the second one is about the Unstructured data,And the third one is about the semi structured data.So the finance data can be divided into three types according to the division of big data.Structured data means you can set those data into pre defined models and patterns and it can be stored into the relationship database.To summary those data have Solid pattern and limited length.For example, those excels to be fitted in are those structured data.For example, your nationality and your race And your sexual diversion.For those enterprises, the operation data from the inner Information system for those enterprises are mostly standardized and structured.To specifically divide those data the first kind is about the basic system. So as to improve the proficiency of human resources and accounting. The second one is about Information system so as to support the operation strategies and to set those marketing analysis so as to broaden the customers field.Traditional business intelligent system use those data produced in the internal Information system about the operation data.Those structured data comes from statistics. So it is the advantage in those area of statistics and progress for those structured data.And because of the characteristic of Structured data. The analysis of data can follow the routine and rules. For example we can predict the income of next month by some Liner Functions.But we need to know that statistics cannot represent all those Information because it has some damage and maybe we will mislead some results.This is because why we seem to get some reasonable conclusion, but it cannot improve our affairs.The second one is about unstructured data.They cannot be defined with solid lengths or solid patterns.For example, those pictures files, news, power point and records of the companies and some texts the social media and some videos and videos.The high value of the big data is much more important. So the unstructured data are those mine I am purified so we need to purify those data so as to dig out some useful value so as to use in the business layer.Most of the daily communication patterns belongs to the unstructured data so in the area of internet eighty percent of the data can be concluded into the unstructured data and the quantity is still increasing.In the field of artificial Intelligence machine learning and analysis of languages and the recognition of pictures need a huge quantity of structured data so as to develop the job and those data the systems are turning into on structure the system.For enterprises, it is much more important for those on structured data, although those enterprises are still collecting and analysis those data stored in the data warehouse or traditional relationship data basis.But those are structured data stored in the system Such as radios, pictures, music texts and videos are being paid much more attention.Unstructured data usually come from the social media,Mobile computing and sensors.For example, those specific communication records facilities and sensors Information and those destination data and those huge quantity of files and images transported through those files, communication protocols and those data of website texts and click flows and those Information from emails and so on.Because of the diverse resources of the data. So Those diverse aspects from the same event can be lived through diverse kinds of data.For example, we can take the consumers Data.The records of consummation can look to the consumption ability and consumption frequency and consumption interest and so on.Those channels Information can look into the preference of these channels.And the payment Information can look into those payment channels and whether or not those customers are paying on the internet social media or whether or not those consumption is directed under the research engine and so on.Those Information or data shows all aspects of the consumption of consumers.And the third one is about semi structured data.Firstly, the semi structured data has some patterns of structure, but it is because of the unstandardized description of those data,It cannot be patterned.Secondly, because of the on limitation of pattern, those data can flow into the system or to renew itself With freedom so as to describe those Event objectively.It is a little bit complex when Re organizing those semi structured data.For example, every staff has its on Resume.But the resume is different from those basic Information and those content in the resume is different.Some resumes contains education situation only and some resume containing complex Information including working situation marriage situation And check on the situation and so on or something unpredicted Information.So usually we need to store those resume.And semi structured data models Handle with this kind of statistic arrangements so as to summarize diversity of those resumes and concern about those Information really concerned by those systems.For example, to set some Sub table.For example, we set those education situation table, the working situation table and so on and then to set some notes on the main table so as to record those key Information and unknown Information Into the notes.And then those semi structured data is turned into structure the data, and it becomes easy and convenient to make some statistics actions.But the disadvantage is that the state cannot be explode or to be researched for those exploration of data.