Web-users’ behavior prediction using markov model with map reduce

  Prof. Vallabh G Patel,  Prof. Komal D Anadkat

As the usage of internet grows, the amount of data produces also increases by time The main challenge. is to store and analyze the data as it is difficult to handle these large amount of data, which proportionally increase the processing time and cost. All user navigation history is stored on the server in form of web log. To predict the next set of web pages that user may visit based on the user’s previous web visit, we need to analyze the web log files. This prediction process in turns will reduce the user’s web access time and pre fetch the next probable pages. Many prediction models are available like SVM, Neural network, ARM but in this research we used markov model. The main drawback of markov is the training time for markov is very large, which increases the response time. To solve this problem, we have trained the markov model in mapreduce programming model of hadoop framework, as it process the large amount of data parallely. The experimental result shows the improvement in response time without compromising the prediction accuracy.

Keywords- Markov model, N-gram ,hadoop ,log mining

Unique Identification Number - IJEDR2101022
Page Number(s) - 176-179
Pubished in - Volume 9 | Issue 1 | January 2021
DOI (Digital Object Identifier) -   
Publisher - IJEDR (ISSN - 2321-9939)

  Prof. Vallabh G Patel,  Prof. Komal D Anadkat,   "Web-users’ behavior prediction using markov model with map reduce", International Journal of Engineering Development and Research (IJEDR), ISSN:2321-9939, Volume.9, Issue 1, pp.176-179, January 2021, Available at :http://www.ijedr.org/papers/IJEDR2101022.pdf