International Journal of Computer Applications |
Foundation of Computer Science (FCS), NY, USA |
Volume 168 - Number 1 |
Year of Publication: 2017 |
Authors: Abdur Rahman Onik, Mashroor Zaman, Rasif Tahmid Islam, Farzana Yeasmeen Sabah |
10.5120/ijca2017914257 |
Abdur Rahman Onik, Mashroor Zaman, Rasif Tahmid Islam, Farzana Yeasmeen Sabah . Predictive Analysis and Warehousing of Web Log Data. International Journal of Computer Applications. 168, 1 ( Jun 2017), 8-14. DOI=10.5120/ijca2017914257
Data mining is a rapidly progressing and increasingly important sector in data science field. The core of data mining is none other than data warehousing, which is gradually becoming a self-sufficient technology for information, integration and data analysis. It is known that the decision support data model has the physical form of data warehouse and as such through storing a huge, thorough and systematic information decisions can be based upon which enterprises act. In this paper, we have analyzed different log file systems of a proxy server. The statistical representation of the log file data were our priority. Data are co-related with each other in every system. How those data are co-related is one of the object of our study. Here we studied the different processes how the data is managed actually. After that we had to consider closely on the types of data that are used so that valuable information and patterns could be discovered. The dynamic nature of modern distributed environments facilitates source data updates and schema changes, even concurrently, in different data sources. It should also be noted that volume of data increases very rapidly as a result. To address the challenge of analyzing data in an efficient way we developed a data warehouse by using multidimensional model. As a means of further analysis, we used predictive analysis on the data to empower appropriate authorities with making useful and accurate decisions.