The basis of data mining microblogging technology marketing

Micro-blog's own operating mechanism, decided to rely on technical means to implement micro-BO marketing, data mining is an inevitable choice can not be bypassed. Weibo is the most marketing value in addition to the people themselves (for human marketing can not rely entirely on technology, the dialogue between people is the most important), is a large number of rapidly changing data. Compared with other marketing platform, for Bo marketing needs from a large number of incomplete, noisy, fuzzy, random data extract implied in it, people do not know beforehand, but is potentially useful information, and even the psychological characteristics of users of Weibo.

Micro-BO data mining is to be able to effectively collect, process, display micro-broad capacity information. Of course, including user ID, name, province, city, registration time, micro-blog release time, gender, whether it is certified users, friends, number of fans, published micro-blog number, published micro-blog and so on. But these are by no means all, perhaps even the lowest-value part of Weibo's massive information content, all of which are easily accessible through the user interface developed by the microblogging platform. But the fundamental purpose of data mining in Weibo is not to obtain these obvious rigid data, but to refine the data to find a valuable link to the enterprise's marketing.

The first is data filtering, micro-blog There is a huge amount of content, even if the filter may be related to the content of the enterprise, may still be a huge amount of data, a business difficult to digest. At the same time, because of the content of the microblogging and short, but also attached a large number of symbols and network language, which brings trouble for participle. It is hard to be precise enough to filter information about the enterprise from the mass microblogging content. In turn, it is possible to set up a segmentation related to the enterprise itself to find the content of the micro-blog, supplemented by a specific user screening, it is easy to find the micro-blog in relation to the enterprise anomalies related to the content, and to track. Starting from the enterprise itself, to micro-Bo search for matching content and user's way can effectively solve the problem of massive content difficult to digest.

Second, the content of micro-blog is too academics, convenient release, interactive way to bring a major problem is the quality of information, information authenticity difficult to identify. Microblogging users to publish information is not simple, the information sent is not all their own intention. Weibo's social relationship is easy to get involved with users, allowing users to passively release certain messages that are being disseminated in order to become part of a group. Many microblogging users forward a content is not agree with or against, not even from the heart of the willing to forward, just to express that I am a member, or I recognized the publisher, or to the reference. In this case the information and the user is actually cut off, only historical tracking history data can determine whether it is the user's true will, to the original micro-broad more to pass this judgment.

Moreover, with the development of micro-blogging to social network, the information form of micro-blogging platform will become more and more inconsistent, now most of the text is presented in the way, there will be a large number of ordinary users will be diverted to a specific application. Although the application of the underlying and data interfaces are mostly the same, but each application itself to the presentation and information, users in which the flow of the way are inconsistent, it is difficult to achieve unified processing. In addition to the improvement of the data mining algorithm itself, it is more closely tied to the bottom of the microblogging platform to gain more advantages in data acquisition.