Data mining

Data mining

1. What is it?

   • "Search for patterns of interest through trees or classification rules, regression techniques, clustering, modeling sequential, dependencies" (Wang, 1999).
   • Is the analysis of files and logs of transactions.
   • Works at the knowledge level in order to discover patterns, relationships, useful rules, associations, or even exceptions for decision making.

2. Applications

   • There are numerous areas where data mining can be applied, in virtually all human activities that generate data:
     • Trade and banking: segmentation of customers, sales forecast, analysis of risk.
     • Astronomy: identification of new ones stars and galaxies.
     • Non-numerical information retrieval: text mining, web mining, search and image, video, voice and text identification of multimedia databases.
     • Medicine and Pharmacy: diagnosis of diseases and the effectiveness of the treatments.
     • Security and fraud detection: facial recognition, identifications biometrics, access to unauthorized networks.

3. Algorithms

   • Supervised or predictive: pred set the value of an attribute of a data set, known other attributes. From data whose label is known, a relationship is induced between said tag and another series of attributes.
   • Unsupervised or Knowledge Discovery: with these algorithms they discover patterns and trends in current data. The discovery of that information is used to carry carry out actions and get a profit from them.

4. Tools

   • Clmentine de SPSS:
     • Clementine focuses on integration of data mining with other processes and business systems that help deliver predictive intelligence in an efficient time during daily business operations
     • Easily access, prepare and integrate data of all kinds from Web pages and surveys.
   • YALE
     • Created at the university of Dortmund
     • Exible enough for knowledge discovery and mining of data.
     • Since YALE is written entirely in Java, it works on the platforms or systems best known operatives

5. Techniques

   • Data Warehousing
     • Large organization process amounts of data of various kinds
     • Facilitates the information retrieval for analytical purposes.
   • Exploratory Data Analysis (EDA)
     • They have a strong computational component ranging from simple statistical methods to the most advanced
     • They aim to determine the relationships between variables when the nature of these relationships is not fully defined

6. Bibliografia You can find more in- https://www.redalyc.org/articulo.oa?id=378343637009 and https://www.redalyc.org/articulo.oa?id=92502902

7. Xelena Castellar, Sebastián Arias

Imágenes y archivos del mapa