Database Management System is a well designed data structure with a set of rules and constraints that allows efficient manipulation and storage of data to ensure their correctness or integrity.
In 2008, Barbara et al, states that DBMS is based on two principles: a three-level conceptual model and alternative models for data organizations.
The three-Level Database Model: Level 1 is conceptual level which contains various user views of the corporate data. Level 2 is logical data level encompassing all organization’s relevant data under data administrator. Level 3 is storage level which specifies the way data is physically stored. Also from material et al, specifies that advantage of the model in individual application programs in Level 1 need not be changed when the physical layer changes.
The four Data model in Business described by Barbara et al, (2008): Hierarchical model “each data element is subordinate to another in a strict manner, like the boxes on an organization chart”, second model is Network model which provides more flexible, less hierarchical approach to data storage used. Each record having multiple parent and child record which forms lattice structure. Third model is Relational model where data stored in tables like spreadsheets representing entities which improves efficiency. This system increases flexibility however, it is not as efficient as previous two models, where navigational maps through the data are predefined. Last model is Object-oriented database model which can be used to store any type of data, spreadsheets, video clips, voice annotation, a photograph or a music segment. Reading material et.al states “Records in an object-oriented model are frequently accompanied by methods that can perform work on the data, and attributes describing the data. The tenets of objects have become increasingly important in the world of Web Services computing, because XML (i.e., HTML) modules utilize object principles” from which one can think objects as black box. Human Genome project is example of such a model.
Early in 1970s organization paid attention to data administration function to manage computerized resourced data as there was data inconsistency from application to application, department to department, site to site, and division to division. As years passed DBMS developed to meet variety of specific tasks, some using mainframe computers while others organization using hundreds to thousands of desktop computers which makes it difficult to manage databases. So organization needed two additional thrusts: Data administrator and data dictionary. (Barbara,2008)
Application software needs database, relational database is fed into it, (Barbara,2008, p.259) states “Enterprise Resource planning is a multi-mode application software seeks to integrate all data and business process into a corporate-wide unified system.”
Data Warehouse is a technology for managing record-based information, where data are generally obtained periodically from transactional database example is “customer data, used to discover how to effectively market current customer as well as noncustomers with same characteristics”. (Reading material,2008)
Key concepts of Data warehouse is: 1. Metadata – It means “data about data” explaining each data element, how it is related to other element, who owns each element, source of each element, who can access each element, and so on. 2. Quality data – Keeping data up-to-date by discarding older data’s. 3. Data marts – “Subset of data pulled off the warehouse for specific user group.” (Barbara,2008)
For building a successful Data-Warehousing project following steps needs to be considered:
1. “To define business used of data.
2. Data model must be created
3. Data must be cleansed.
4. The tools for the analysts must be selected, and
5. Usage and data performance must be monitored.”(Reading Material)
It is the area that deals with analysis of large and complex databases to discover new, useful and interesting knowledge using techniques from statistics and machine learning.
Application of data mining and artificial intelligence are successful in many domains, one such domain is in Medical field, (Fariba,2008) RIDC-ANNE is a novel algorithm developed that tries to handle complex clinical data(data mining) and rule extraction from individual black box model. This approach differentiates between successful and unsuccessful kidney transplants. It uses trail and error approach to determine appropriate network topology automatically, resulting 87% accuracy with average data sets. (Fariba,2008 p.4) “Perhaps more success stories in data mining are due to advances in computer science and technology. From these success stories, we have learned that the use of artificial intelligent and data mining for medicine can be very effective and successful.”