• Subba Rao Peram
• B Pramamayudu
• B Tarakeswara Rao
• E Deepak Chowdary

PC framework techniques have extended in intricacy to where manual assessments of framework conduct for reasons for abnormality recognition have wound up being inconceivable. As these frameworks result enormous logs of their errand, gear drove examination of them is an extending need with effectively various existing arrangements. These to a great extent rely upon having hand-created highlights, call for crude log preprocessing and include evacuation or utilize observed finding requesting having really a named log database not in every case helpfully available. We propose a two section profound auto encoder LSTM model gadgets which may require the no high quality ascribes, no preprocessing of the data which is manages crude message just as yields a peculiarity groove for each and every log access. In this peculiarity rating denotes the uncommonness log occurrenceof both as far as its substance and furthermore transient setting. This was prepared just as analyzed logs of HDFS including two million crude lines out of this 50% was utilized to preparing just as 50% for testing. While this model can't coordinate with the exhibition of a directed parallel classifier, it very well may be a gainful apparatus as an unrefined channel for hand-worked assessment of log archives where a recognized database is blocked off.

"Anomaly Detection From Log Data With Long-Short Term Memory Network", International Journal of Emerging Technologies and Innovative Research (www.jetir.org), ISSN:2349-5162, Vol.8, Issue 5, page no.f185-f188, May-2021, Available :http://www.jetir.org/papers/JETIR2105690.pdf

"Anomaly Detection From Log Data With Long-Short Term Memory Network", International Journal of Emerging Technologies and Innovative Research (www.jetir.org | UGC and issn Approved), ISSN:2349-5162, Vol.8, Issue 5, page no. ppf185-f188, May-2021, Available at : http://www.jetir.org/papers/JETIR2105690.pdf