Data Tagging for Data Quality Modeling - 1 of ?

Many similarities exist between producing quality data and manufacturing quality products. Similarities such as quality measures, conformity to specifications, low defect rate, and improved user satisfaction.

If we go back to the development of the machine and the industrial mass production, we will find similarities to the advancement of the data processing power and the mass production of large volume of data (stored in data warehouses) that is widely distributed and easily accessed.

Faced by the reality of advancing data processing power, several organizations are opted or forced to manage large volume of data. Some organizations are motivated by the relatively cheap technology enabled their vision. Others are scared of the competition and the demand from data consumers for different views of data.

In reality, the standards, specifications, the engineering methods, and the personnel with the know-how are not advancing in parallel. This situation led to chaos in the quality of data presented to the data consumers in specially in the business or government operation paradigm.

This post will discuss the data quality within the context of application and data integration, which as argued has played a big role to uncover data quality chaos. I will examine the following aspects in data quality:
1) The current issues in data quality.
2) What is meant by quality data.
3) XML specifications and modeling techniques as a tagging solution to manage and control data quality.

Using Data Mining Techniques For Fraud Control - 1 of ?

Fraud is defined as a deliberate use of deception to conduct illicit activities. Recently, fraud has become a serious and a rising problem as the result of the exponential growth in the electronic transactions. This post has argued that over the last ten years, to be specific, a major shift occurred in the way transactions are initiated and fulfilled. This shift has created a new type of electronic fraud including but not limited to misrepresentation, identity theft, and electronic hacking. So far, organizations and companies continue to rely on antiquated manual processes to detect unauthentic or suspicious transactions.

This post is an attempt to demonstrate the untapped potentials of data mining techniques as tool to discover hidden knowledge in all sort of transactions with purpose to control and detect fraud. I will examine the emerging technologies that have enabled the advancement of data mining to the front lines of information science. I will review the progress in software and hardware technologies then discussed the contribution of data management technologies to the way data collected, stored, and accessed. I will take a closer look at the evolution of the way data is organized, stored, and discovered. In this context I will examine database management systems and data modeling design.

It argued that the expansion of using electronic transaction and data storage in large enterprise wide databases and advancement in the area of data warehousing have brought heterogeneous data into one global data model which made it easier to apply data mining algorithm with little effort in data preparation. Also, the paper has argued the materialization of data mining techniques is a result of the convergence of the development in the computer power and the improvement in data collection and storage.

In future posts, if I get the time, I will review in details several data mining techniques and algorithms. I will also, discuss the argument that despite the advancement of data mining techniques and the promising results, both the techniques and the results sometime are overstated or hyped. However, the future of data mining application is limitless and there are numerous commercial applications that are my top candidate for embedded data mining model such as database management systems and Enterprise Resource Planning (ERP) application.