10 Hits in 0.96 sec

Problem Identification by Mining Trouble Tickets

Vikrant Shimpi, Maitreya Natu, Vaishali P. Sadaphal, Vaishali Kulkarni
2014 International Conference on Management of Data  
= Mode of the position of the word w * Label of the clique = Arrangment of the words in Cw according to p Figure 2 : 5 . 25 Figure 2: System-generated tickets: group id vs size of each group Figure  ...  a label • Identify set of common words Cw from the set of cleaned description belonging to the clique -For each word w in Cw, * Compute its position in set of cleaned description belonging to clique, p  ... 
dblp:conf/comad/ShimpiNSK14 fatcat:b56slvpiofdrffjzcvmdp5cfom

Sensor Selection Heuristic in Sensor Networks [chapter]

Vaishali P. Sadaphal, Bijendra N. Jain
2005 Lecture Notes in Computer Science  
We consider the problem of estimating the location of a moving target in a 2-D plane. In this paper, we focus attention on selecting an appropriate 3 rd sensor, given two sensors, with a view to minimize the estimation error. Only the selected sensors need to measure distance to the target and communicate the same to the central "tracker". This minimizes bandwidth and energy consumed in measurement and communication while achieving near minimum estimation error. In this paper, we have proposed
more » ... hat the 3 rd sensor be selected based on three measures viz. (a) collinearity, (b) deviation from the ideal direction in which the sensor should be selected, and (c) proximity of the sensor from the target. We assume that the measurements are subject to multiplicative error. Further, we use least square error estimation technique to estimate the target location. Simulation results show that using the proposed algorithm it is possible to achieve near minimum error in target location. the central device responsible for estimating the location of the target, (b) the clocks of the sensors are synchronized so that the sensors "measure" distance at approximately the same time, and (c) the sensors are able to communicate their measurements to this central device, also referred to as the "tracker". Since the target is moving and since a sensor must be within a certain distance from the target (before it can detect the presence of the target and measure distance), we assume that there are several sensors, {s i } = Σ, spread across the 2-D plane. In fact, we assume that there are three or more sensors located in and around every point in the 2-D plane so that we can compute an estimate based on measurements from a subset of three sensors suitably selected to minimize estimation error. This approach also allows one to minimize communication overheads and conserve battery power available to sensors. Further, since the target is moving, the collection of sensors changes every time an estimate is required to be obtained. Specifically, we assume that as the target moves, if sensors {s 1 , s 2 , s 3 } have made measurements at time t k , then at time t k+1 , we drop one of the sensors s 1 , s 2 , or s 3 and select a sensor s 4 suitably so as to minimize the error in estimated location of the target. Accordingly, this paper is about suitably selecting the 3 rd sensor from a set of N k+1 sensors.
doi:10.1007/11602569_23 fatcat:63k6vlf3w5fqpdp6fygsy7cnda

Analyzing Periodically Occurring Patterns in Time Series

Shivam Sahai, Maitreya Natu, Vaishali P. Sadaphal
2010 International Conference on Management of Data  
Thus, ∀(mi, mj ) ∈ LM such that T ime(mi) < T ime(mj ): pair (mi, mj ) ∈ LM P if, (p − δ) ≤ (T ime(mj ) − T ime(mi) ≤ (p + δ) This ensures that ∀(m i , m j ) ∈ LM P , the pattern T p will be bound by minima  ...  . • Time-series region: A time-series region T p of length p is a subsequence of p contiguous points in the timeseries.  ... 
dblp:conf/comad/SahaiNS10 fatcat:kqvpdyd2ejbdfmvpkolyaihcza

Random and Periodic sleep schedules for target detection in sensor networks

Vaishali P. Sadaphal, Bijendra N. Jain
2007 2007 IEEE Internatonal Conference on Mobile Adhoc and Sensor Systems  
Specifically, we analyse and obtain for the Random wake-up schedule the expected delay in detection, and the delay, such that with probability P, the delay is less than the computed value.  ...  As a result ∃p a q a n = p a m + a. For sure p a ≥ 0.  ...  Then, ε = δ s − δ t = 1E(∆) vs. δ s . and p 1 = 0.1037, p = 0.1, p 2 = 0.0906, p 3 = 0.08, etc.  ... 
doi:10.1109/mobhoc.2007.4428677 dblp:conf/mass/SadaphalJ07 fatcat:sk7zscf4wzdezm3o556vsjwl4a

Table of Contents PROYEVA: System to Evaluate the Projects Quality in Contests Community-Commerce Brokering Arena for Opportunistic Cloud Services Offerings An Approach to Find Integration and Monitoring Points for Container Logistics Business Processes A Monitoring Approach for Dynamic Service-Oriented Architecture Systems

Laura Silvia, Vargas Perez, Agustin Francisco Gutierrez-Tornes, Edgardo Felipe-Riveron, Ethan Hadar, Steven Greenspan, Tugkan Tuglular, Dilek Avci, Sevket Cetin, Gokhan Daghan, Murat Ozemre, Tolgahan Oysal (+8 others)
Lopez-Soler Workload Characterization for Stability-As-A-Service Vaishali P.  ...  Sadaphal and Maitreya Natu 84 4R of Service Innovation: Research, Requirements, Reliability and Responsibility Anastasiya Yurchyshyna, Abdelaziz Khadraoui, and Michel Leonard Visualization Method Based  ... 

Analytics-Based Solutions for Improving Alert Management Service for Enterprise Systems

Anuja Kelkar, Utkarsh Naiknaware, Sachin Sukhlecha, Ashish Sanadhya, Maitreya Natu, Vaishali Sadaphal
2013 2013 IEEE 13th International Conference on Data Mining Workshops  
Sunday and Monday. 2) p (dow=Sunday) = 1/16 = 0.06; 3) p (dow=M onday) = 15/16 = 0.93; 4) We next compute the entropy value for the day-of-week dimension: H dow = i∈Sunday,M onday −p i * log(p i ). 5)  ...  Given a set with n possible values {x 1 , x 2 , . . . x n }, the entropy is defined as H = n i −p i * log(p i ), where p i is the probability of occurrence of the value x i .  ... 
doi:10.1109/icdmw.2013.166 dblp:conf/icdm/KelkarNSSNS13 fatcat:twnr3okqjzemlodfrgkskcqmwi

Varanus: More-with-less fault localization in data centers

Vaishali Sadaphal, Maitreya Natu, Harrick Vin, Prashant Shenoy
2012 2012 Fourth International Conference on Communication Systems and Networks (COMSNETS 2012)  
Formally, information entropy H m (T ) for each candidate monitor node m ∈ T is defined as: H m (T ) = −p(T R m )log(p(T R m )) − p(T U m )log(p(T U m )) where p(T R m ) and p(T U m ) , respectively, denote  ...  The p-value thus obtained represents the expected amount of similarity between two current observation windows. We use this p-value as the similarity threshold.  ... 
doi:10.1109/comsnets.2012.6151303 dblp:conf/comsnets/SadaphalNVS12 fatcat:qkdu4ditkrbz3dplmqbpssccea

Predico: A System for What-if Analysis in Complex Data Center Applications [chapter]

Rahul Singh, Prashant Shenoy, Maitreya Natu, Vaishali Sadaphal, Harrick Vin
2011 Lecture Notes in Computer Science  
We model each node as a M/G/1/P S queue i.e. the service times are assumed to have an arbitrary distribution and the service discipline at each node is assumed to be processor sharing (PS).  ... 
doi:10.1007/978-3-642-25821-3_7 fatcat:vvfdhrnhereq7o6tgzfdxnw5pq

Analytical modeling for what-if analysis in complex cloud computing applications

Rahul Singh, Prashant Shenoy, Maitreya Natu, Vaishali Sadaphal, Harrick Vin
2013 Performance Evaluation Review  
We model each node as a M/G/1/ P S queue i.e. the service times are assumed to have an arbitrary distribution and the service discipline at each node is assumed to be processor sharing (PS).  ... 
doi:10.1145/2479942.2479949 fatcat:kmazmz6farasheawfgcw7ln3x4

Workload Characterization for Stability-As-A-Service

Vaishali Sadaphal, Maitreya Natu
Then the p-prediction heuristic states that if (p 12 < M p ) then p 3 > min(p 1 , p 2 ), where M p is the threshold defined for the p-prediction heuristic.  ...  Let the p-values of the t-test ran on performance data of these subsets and that of the rest of data are p 1 and p 2 respectively.  ...