Hidden Markov Models

The main objective this study is to propose development of a new subcarrier-mapping scheme using GIFDMA
for the LTE/LTE-A system. This new scheme addresses the user and frequency diversity by allocating the
subcarrier and combining the best features of the traditional IFDMA and LFDMA schemes. The Interleaved
FDMA (IFDMA) is a special case of GIFDMA that uses non-contiguintoous subcarrier in contrast to LFDMA
that utilizes adjacent subcarrier. Both LFDMA and IFDMA subcarrier mappings significantly reduce PAPR at
the expense of high time complexity, in order to reduce the time complexity and heuristic correlation between
interleave level and roll-off factors for fast PAPR estimates, a Multilayer Perceptron Artificial Neural Network
(MLP-ANN) and Random Forest (RF) algorithms are introduced. RF is a Supervised Machine Learning Algorithm that is used widely in Classification and Regression problems. It builds decision trees on different samples and takes their majority vote for classification and average in case of regression.

Message passing algorithms may be viewed from a purely probabilistic or statistical physics perspective. This works describes an alternative, linear algebraic view of message passing in trees. We demonstrate the construction of global belief operators on Markov Chains and Trees and compare these with classical results. By interpreting message passing as finding a global stable solution we wish to avoid application dependent development of message passing methods of more complex graph structures.

Markov Process are one stochastik process where the future not only one term that influence present that was known. Hidden Markov Mod el is defined as five parameters (N,M,A,B, π). If λ = {A, B,π} are variable parameters. N and M are HMM fixed pa rameter. Hidden Markov Model are developed from markov chain where term and present situation not only determined by currently situation from the ano ther term and future situation but also sequence change. The sequence are one hidden part f rom Hidden Markov Model. One kind of Stock Market Forecasting will concern on available s quence then applied it on the Hidden Markov Model. In this research, using data that cam e from stock market which come from “Yahoo! Finance” in the web “http://finance.yahoo.c m”, that will result the interpretation of stock market movement. Key word : Forecasting, Hidden Markov Model, StockMarket Movement I. PENDAHULUAN Peramalan (forecasting) merupakan salah satu bagian vital bagi setiap organisasi bisnis dalam ...

In this paper, the aim is to investigate weather or not that changing the filter-bank components (of the speaker recognition system) could improve the system performance in identifying the speaker. The filter is composed of 30 Gamatone filter channels. First, the channels are mel distributed of the frequency line. Then the componentsvalues (center frequencies and bandwidths) changes with each run. Genetic algorithm (GA) is adopted to improve the filter component values that, in a result, improve the system performance. At each GA run, a new set of filter components will be generated that aimed to improve the performance comparing with the previous run. This will continue until the system reach to the maximum accuracy or the GA reach to its limits. Results show that the system will be improved at each run, however, different words might response differently to the system filter changing. Also, in terms of additive noise, the results show that although the digits affected differently by the noise, the system still get improving with reach GA run.

With the growing usage of computer systems in daily life, a natural and intuitive Human Computer Interaction (HCI) method to support the embedding of computer systems in our environment seems necessary. Gestures are of utmost importance for the design of natural user interfaces. Hand gesture recognition to extract meaningful expressions from the human hand movements and postures is being used for different applications. However, the recognition of hand gestures that contain different hand poses can be challenging. In this ...

Designers of fault-tolerant computer systems need methodological and software framework which would support their efforts in analysis and optimization of new design solutions, based on new and forthcoming hardware and software technologies, embedded systems, in particular These new and advanced technologies-high-performance and self-reconfigurable systems, nanotechnologies-lead to unprecedented challenges. For example, often as a result of transient faults, reconfigurations in FPGA-based high-performance ...

End-to-end neural networks have shown promising results on large vocabulary continuous speech recognition (LVCSR) systems. However, it is challenging to integrate domain knowledge into such systems. Specifically, articulatory features (AFs) which are inspired by the human speech production mechanism can help in speech recognition. This paper presents two approaches to incorporate domain knowledge into end-to-end training: (a) fine-tuning networks which reuse hidden layer representations of AF extractors as input for ASR tasks; (b) progressive networks which combine articulatory knowledge by lateral connections from AF extractors. We evaluate the proposed approaches on the speech Wall Street Journal corpus and test on the eval92 standard evaluation dataset. Results show that both fine-tuning and progressive networks can integrate articulatory information into end-to-end learning and outperform previous systems.

Kanuri is a Nilo-Saharan language spoken in the Lake Chad basin of West and Central Africa. Effective part-of-speech (POS) tagging is crucial for natural language processing tasks in Kanuri, such as machine translation, information extraction, and text generation. However, the lack of comprehensive linguistic resources and annotated datasets for Kanuri has hindered the development of accurate POS taggers for this language. This study aims to develop a part-of-speech tagger for the Kanuri language using a Hidden Markov Model (HMM) approach. The goal is to create a robust and accurate POS tagging system that can be used to support various natural language processing applications for the Kanuri language. The study involved a development of corpus of Kanuri text collected from various sources. A HMM-based POS tagging model was designed and trained on the annotated Kanuri corpus. The HMM-based POS tagger achieved an overall accuracy of accuracy of 0.827 % on the Kanuri test data. The developed HMM-based POS tagger can be integrated into various natural language processing pipelines for Kanuri, enabling more advanced language analysis and understanding tasks. Additionally, the annotated Kanuri corpus can be used to expand and improve the POS tagging model further and support the development of other language technologies for the Kanuri language.

Statistical modelling techniques for automatic reading systems substantially rely on the availability of compact and meaningful feature representations. State-ofthe-art feature extraction for offline handwriting recognition is usually based on heuristic approaches that describe either basic geometric properties or statistical distributions of raw pixel values. Working well on average, still fundamental insights into the nature of handwriting are desired. In this paper we present a novel approach for the automatic extraction of appearance-based representations of offline handwriting data. Given the framework of deep belief networks – Restricted Boltzmann Machines – a two-stage method for feature learning and optimization is developed. Given two standard corpora of both Arabic and Roman handwriting data it is demonstrated across script boundaries, that automatically learned features achieve recognition results comparable to state-of-theart handcrafted features. Given these promising r...

In this paper, we present the analysis of GlobalPhone (GP) and speech corpora of Ethiopian languages (Amharic, Tigrigna, Oromo and Wolaytta). The aim of the analysis is to select speech data from GP for the development of multilingual Automatic Speech Recognition (ASR) system for the Ethiopian languages. To this end, phonetic overlaps among GP and Ethiopian languages have been analyzed. The result of our analysis shows that there is much phonetic overlap among Ethiopian languages although they are from three different language families. From GP, Turkish, Uyghur and Croatian are found to have much overlap with the Ethiopian languages. Moreover, morphological complexity of the GP and Ethiopian languages, reflected by type to token ration (TTR) and out of vocabulary (OOV) rate, has been analyzed. Korean and Amharic have been identified as extremely morphologically complex compared to the other languages. Tigrigna, Russian, Turkish and Polish are also among the morphologically complex l...

Voice quality is currently a key issue in speech synthesis research. The lack of realistic intra-speaker voice quality variation is an important source of concern for concatenation-based synthesis methods. A challenging problem is to reproduce the voice quality changes that are occuring in natural speech when the vocal e ort is varying. A new method for voice quality modi cation is presented. It takes advantage of a spectral theory for voice source signal representation. An algorithm based on periodic-aperiodic decomposition and spectral processing (using the short-term Fourier transform) is described. The use of adaptive inverse ltering in this framework is also discussed. Applications of this algorithm may include: pre-processing of speech corpora, modi cation of voice quality parameters together with intonation in synthesis, voice transformation. Some experiments are reported, showing convincing voice quality modi cations for various speakers.