Markov Chain Learning on File Access Patterns with Noisy Data
نویسنده
چکیده
File access patterns for application startup are fixed and predictable. More precisely, they would obey the Markovian property of each future access depending only on the current access. In this project, we attempt to learn the Markov chain transition probabilities, where each file access is a state in the chain. But since multiple applications may run at the same time, the file access chains for one process would have noise from the other processes. We attempt to filter out noise with different estimated threshold values and show the trade-offs with each value.
منابع مشابه
Space-Optimized Markov Chain Model for File Prefetching
This project investigated the ability of a Markov Chain model to predict le access patterns in OceanStore, a global-scale storage system currently under development. Because a na ve implementation of the transition matrix is an ineecient use of memory , we evaluated a simple sparse-matrix technique as a space optimization.
متن کاملLearning Hybrid Bayesian Networks by MML
We use a Markov Chain Monte Carlo (MCMC) MML algorithm to learn hybrid Bayesian networks from observational data. Hybrid networks represent local structure, using conditional probability tables (CPT), logit models, decision trees or hybrid models, i.e., combinations of the three. We compare this method with alternative local structure learning algorithms using the MDL and BDe metrics. Results a...
متن کاملNovel machine learning techniques for anomaly intrusion detection
Novel machine learning techniques for anomaly intrusion detection" (2004). ABSTRACT This paper explores the methodology of using kernels and Support Vector Machine (SVM) for intrusion detection. A new insight into two well known anomaly detection algorithms-STIDE and Markov Chain anomaly detectors, is achieved using kernel theory. We introduce two new classes of kernels used for intrusion detec...
متن کاملBAT - The Bayesian analysis toolkit
We describe the development of a new toolkit for data analysis. The analysis package is based on Bayes’ Theorem, and is realized with the use of Markov Chain Monte Carlo. This gives access to the full posterior probability distribution. Parameter estimation, limit setting and uncertainty propagation are implemented in a straightforward manner. A goodness-of-fit criterion is presented which is i...
متن کامل