Brief Review of Data Mining
Data Mining is an integral part of data analysis that contains a series of activities to “sense” ideas, “analyze” data and “interpretation” and “assessment” resulted. Various stages of the technique are:
Objectives Analysis: Sometimes it is very difficult to define statistically the phenomenon we want to analyze. In fact, the company’s objectives are often unclear, but may be difficult to formalize. A clear understanding of the crisis and objectives of installation is very important to analyze correctly. This is probably one of the most complex processes, because it establishes procedures for use and, as such objectives should be clear and should not be any doubt or ambiguity.
Group data collection and preprocessing: Once goals are set and defined analysis, we need to collect the necessary data or choose to study. First, it is essential to recognize data sources. Usually, data is collected from domestic sources, is cheaper and more reliable and, moreover, these data also have the advantage of being the result of experience and company procedures.
Analysis of survey data and conversion: This step includes a preliminary analysis of available information. This is a preliminary assessment of the importance of data collected. Exploratory analysis and / or research may reveal irregular data. An exploratory analysis is important because it allows the analyst to choose the most appropriate method for statistical analysis.
The choice of statistical methods: There are several statistical methods can be used for analysis, it is important to classify the existing methods. Statistical method of choice is specific and depends on the cause of the problem and also the type of information available.
Data analysis based on selected methods: Since the statistical method chosen, it must be translated into algorithms suitable for operating results. The range of skilled and unskilled software are widely available for data mining and as such is not always necessary to develop algorithms for ad-hoc “standard”. However, it is essential that the management method of data mining and knowledge to have a good knowledge and understanding of different methods of data analysis and various software solutions available, so it can adapt well when needed, and the company can fully interpret the results.
Evaluation and contrast techniques and selection of final exam: It is necessary to choose the best “model” the variety of available statistical methods. The choice should be based model contrasts with the results. In assessing the performance of a specific statistical method and / or type, all dependents and / or criteria should also be considered. Criteria may be other restrictions on society, both in terms of time and resources or, perhaps, in terms of quality and accessibility of data.
Elucidation statistical model chosen and its use in decision making: the field of data mining is not limited to data analysis and is not also include the integration of results to facilitate business decision making. Entrepreneurial awareness, building standards and their use in decision-making allows us to advance in the diagnostic phase of decision-making phase. Once the model is completed and tested a set of information classification rule can be generalized. But the inclusion process of data mining in business should not be made hastily, but should always be done slowly, the institution is sound and logical. The ultimate goal of data mining is an integral part of business decisions.
No related posts.