Current Trends in Data Mining
Data mining is based on complex algorithms that allow for the segmentation of data to identify patterns and trends, detect anomalies, and predict the probability of various situational outcomes. The raw data being mined may come in both digital and analog formats depending on the data sources. What is most important about data mining is the quality and amount of the data, regardless of the format.
With the amount of data being collected and stored today, data mining has taken on much greater significance for public and private organizations of all types. For example:
- Law enforcement agencies can use data mining to look for crime patterns.
- Retailers can predict buyer behavior and shopping trends, as well as analyze marketing results and customer loyalty patterns.
- Financial firms can identify loan repayment patterns and any outlying activities that might point to fraud.
- Medical researchers can uncover biological markers that lead to new discoveries in the treatment of disease.
- Telecommunications providers can identify unusual calling patterns and potentially fraudulent activity.
- Scientists can more quickly apply their hypotheses by accessing data collected during previous experiments.
Even fields that have been slow to adopt “Big Data” are on the bandwagon, extracting vital business intelligence through data mining. As the discipline continues to grow in sophistication, new trends have emerged:
Distributed Data Mining (DDM)
Distributed data mining allows for access to volumes of data that are housed at several different company sites or at various organizations. Highly sophisticated algorithms are created to retrieve the necessary data regardless of where it is stored so that it can be applied to a specific data model that will deliver the proper knowledge and reports.
Multimedia Data Mining
Multimedia data mining is just what it sounds like. The practice looks to extract relevant data from text, hypertext, audio, video, still images and other content, and then convert that data into a numerical representation of knowledge. Multimedia data mining can be used to identify associations, clustering and classification, as well as perform similarity search.
Spatial and Geographic Data Mining
This type of data mining concerns itself with astronomical, environmental and geographical data, including aerial images captured from space. Spatial and geographic data mining can indicate measures such as distance and topology, which can be used for navigation and geographic information systems (GIS) applications.
Ubiquitous Data Mining
Ubiquitous data mining (UDM) involves tapping into mobile devices to access data about individuals. There are many challenges with this type of data mining including cost, complexity, and privacy. Yet, the opportunities for presented for business and industry are enormous, especially for studies into human-computer interaction.
Time Series and Sequence Data Mining
The study of cyclical and seasonal trends is a primary application of time series and sequence data mining, although it is also useful in analyzing random events outside of an expected sequence. This practice is especially useful for retail companies in terms of assessing customer behaviors and buying patterns.
Sources:http://docs.oracle.com/cd/B28359_01/datamine.111/b28129/process.htm#DMCON002 http://www.slideshare.net/dataminingtools/data-mining-application-and-trends-in-data-mining http://www.slideshare.net/IJMER/es2646574663