China
Africa
Explore chapters and articles related to this topic
Recent Advancements in Automatic Sign Language Recognition (SLR)
Published in Sourav De, Paramartha Dutta, Computational Intelligence for Human Action Recognition, 2020
Varshini Prakash, B.K. Tripathy
A forward backward algorithm is used to generate the probability P(O|λ) where π denotes the initial state probabilities, a denotes the transition state probabilities and b denotes the output probabilities. α1(i)=πibi(O1)∀i if i∈SI,πi=1nI else πi=0
Non-Gaussian and Nonlinear State Space Models
Published in Randal Douc, Eric Moulines, David S. Stoffer, Nonlinear Time Series, 2014
Randal Douc, Eric Moulines, David S. Stoffer
This recursion, summarized in Algorithm 9.2, is the forward-backward algorithm or the Baum-Welch algorithm for discrete Hidden Markov Models. In the forward pass, the filtering distributions ϕξ,t,t∈{0,…,n} are computed and stored. In the backward pass, these filtering distributions are corrected by recursively applying the backward kernels.
Markov Model Inferencing in Distributed Systems
Published in S. Sitharama Iyengar, Richard R. Brooks, Distributed Sensor Networks, 2016
Chen Lu, Jason M. Schwier, Richard R. Brooks, Christopher Griffin, Satish Bukkapatnam
The Baum–Welch algorithm makes use of the forward–backward algorithm to estimate the transition matrix P of a HMM [1,24]. In addition to the observed sequence of symbols y, the algorithm requires an initial estimate of the transition structure (in the form of a guess for P and the initial probability distribution across the states π0) be available. As a result, this algorithm requires initial knowledge of the structure of the Markov process governing the dynamics of the system producing the output. The algorithm is iterative and the stopping criteria may be given in terms of the convergence of the solution in any number of metrics.
The key modules involved in the evolution of an effective instrumentation and communication network in smart grids: a review
Published in Smart Science, 2023
Some new terms and acronyms were used in this paper. Distributed control station (DCS) is a system having more than one process control station with different input output units. Advanced metering infrastructure (AMI) is a new technology in smart grid which consists of meters, two-way communication, and data management system [5]. Energy management system (EMS) is a system that contains computer-aided tools to monitor and control the performance the generation, transmission, and consumption. Distributed energy resources (DER) are power generation resources located near load centers. Forward–backward algorithm is an algorithm that calculates posterior marginals of all hidden state variables given as a sequence of observations. Weighted average price prediction is an estimation algorithm which calculates the price of electricity. Load shifting is the ability to store the battery power and sell it back to the grid. Distributed and hierarchical control are the two most widely used control methods. In hierarchical control, each controllable object in the grid is controlled by a local controller which is eventually controlled by a central controller. Distributed controllers control individually and communicate each other to attain global goals. 4GLTE is a particular type of faster mobile internet communication.
A Machine Learning Approach for Quantifying the Design Error Propagation in Safety Critical Software System
Published in IETE Journal of Research, 2022
Therefore the reliability predictions rely on effects of faults, which turn them into errors. In safety critical software system (SCSS), the failures are predicted through the analysis of software error patterns the system is experiencing. Since the software error states are unknown and cannot be measured directly, it is considered as hidden states. Hence we are motivated to adopt the HMM algorithm [18] which can be implemented in two main steps: firstly to fix the number of hidden states and secondly to compute the probability that leads to failure. Once the system is modeled and the temporal distributions are determined, HMM helps traverse the states using forward–backward algorithm. This practice has been applied to several complex systems, starting from speech recognition [18], DNA and protein evolution, failure prediction [19] and so on.
Functional Uncertainty Analysis of Phasor Measurement Unit Using Fuzzy Hidden Markov Model
Published in IETE Journal of Research, 2018
Soumita Ghosh, Debomita Ghosh, Dusmanta Kumar Mohanta
The probability of occurrence of an observation sequence {Ob1,…,ObT}, given model , can be evaluated by the forward–backward algorithm. The forward and backward prediction variables can be expressed in terms of fuzzy as detailed in Equations (2)–(4). Let represent the grade of certainty of the chosen observation and state Sti at time “t”.