COMBINED FEATURE EXTRACTION TECHNIQUES AND NAIVE BAYES CLASSIFIER FOR SPEECH RECOGNITION

Speech processing and consequent recognition are important areas of Digital Signal Processing since speech allows people to communicate more natu-rally and efficiently. In this work, a speech recognition system is developed for re-cognizing digits in Malayalam. For recognizing speech, features are to be ex-tracted from speech and hence feature extraction method plays an important role in speech recognition. Here, front end processing for extracting the features is per-formed using two wavelet based methods namely Discrete Wavelet Transforms (DWT) and Wavelet Packet Decomposition (WPD). Naive Bayes classifier is used for classification purpose. After classification using Naive Bayes classifier, DWT produced a recognition accuracy of 83.5% and WPD produced an accuracy of 80.7%. This paper is intended to devise a new feature extraction method which produces improvements in the recognition accuracy. So, a new method called Dis-crete Wavelet Packet Decomposition (DWPD) is introduced which utilizes the hy-brid features of both DWT and WPD. The performance of this new approach is evaluated and it produced an improved recognition accuracy of 86.2% along with Naive Bayes classifier.

An Effective Path Protection Method to Attain the Route Stability in MANET

With the increasing popularity of wireless network and its application, mobile ad-hoc networks (MANETS) emerged recently. MANET topology is highly dynamic in nature and nodes are highly mobile so that the rate of link failure is more in MANET. There is no central control over the nodes and the control is distributed among nodes and they can act as either router or source. MANTEs have been considered as isolated stand-alone network. Node can add or remove at any time and it is not infrastructure dependent. So at any time at any where the network can setup and a trouble free communication is possible. Due to more chances of link failures, collisions and transmission errors in MANET, the maintenance of network became costly. As per the study more frequent link failures became an important aspect of diminishing the performance of the network and also it is not predictable. The main objective of this paper is to study the route instability in AODV protocol and suggest a solution for improvement. This paper proposes a new approach to reduce the route failure by storing the alternate route in the intermediate nodes. In this algorithm intermediate nodes are also involved in the route discovery process. This reduces the route establishment overhead as well as the time to find the reroute when a link failure occurs.

Biochemical and physiological characteristics of actinomycetes isolated from high altitude shola soils of tropical Montane forest

Actinomycetes are gram-positive, free-living, saprophytic bacteria widely distributed in soil, water and colonizing plants showing marked chemical and morphological diversity. They are potential source of many bioactive compounds, which have diverse clinical effects and important applications in human medicine. In the present work, we have studied some of the physiological and biochemical characteristics of 36 actinomycete strains isolated from the shola soils of tropical montane forest; a relatively unexplored biodiversity hotspot. Ability of actinomycetes isolates to ferment and produce acids from various carbohydrate sources such as innositol, mannose, sorbitol, galactose, mannitol, xylose, rhamnose, arabinose, lactose and fructose were studied. Almost all the carbon compounds were utilized by one or other actinomycete isolates. The most preferred carbon sources were found to be xylose (94.44%) followed by fructose and mannose (91.66%). Only 41.76% of the isolates were able to ferment lactose. The ability of actinomycetes isolates to decompose protein and amino acid differ considerably. 72.22% of the isolates were able to decompose milk protein casein and 61.11% of the isolates decompose tyrosine. Only 8.33% of the strains were able to decompose amino acid hypoxanthine and none of them were able to decompose amino acid xanthine. Potential of the actinomycetes isolates to reduce esculin, urea and hippurate and to resist lysozyme was also checked. 91.66% of the isolates showed ability to decompose esculin and 63.88% of the isolates had the capacity to produce urease and to decompose urea. Only 25% of the isolate were able to decompose hippurate and 94.44% showed lysozyme resistance

Spatial and Temporal Variation of Microbial Population in the Grassland Soils of Tropical Montane Forest: Influence of Soil Physico-Chemical Characteristics and Nutrients

Present study is focused on the spatiotemporal variation of the microbial population (bacteria, fungus and actinomycetes) in the grassland soils of tropical montane forest and its relation with important soil physico-chemical characteristics and nutrients. Different physico-chemical properties of the soil such as temperature, moisture content, organic carbon, available nitrogen, available phosphorous and available potassium have been studied. Results of the present study revealed that both microbial load and soil characteristics showed spatiotemporal variation. Microbial population of the grassland soils were characterized by high load of bacteria followed by fungus and actinomycetes. Microbial load was high during pre monsoon season, followed by post monsoon and monsoon. The microbial load varied with important soil physico-chemical properties and nutrients. Organic carbon content, available nitrogen and available phosphorous were positively correlated with bacterial load and the correlation is significant at 0.05 and 0.01 levels respectively. Available nitrogen and available phosphorous were positively correlated with fungus at 0.05 level significance. Moisture content was negatively correlated with actinomycetes at 0.01 level of significance. Organic carbon negatively correlated with actinomycetes load at 0.05 level of significance

SPATIOTEMPORAL VARIATION AND ANTIBACTERIAL ACTIVITY OF ACTINOMYCETES ISOLATED FROM HIGH ALTITUDE GRASSLAND SOILS OF TROPICAL MONTANE FOREST – KERALA, INDIA

TThe invention of novel antibiotics and other bioactive microbial metabolites continues to be an important aim in new drug discovery programmes. Actinomycetes have the potential to synthesize lots of diverse biologically vigorous secondary metabolites and in the last decades actinomycetes became the most productive source for antibiotics. Therefore in the present study we analyze the antibacterial activity of the actinomycetes isolated from grassland soil samples of Tropical Montane forest. A total of 33 actinomycete strains isolated were characterized and screened for antibacterial activities using well diffusion method against six specific pathogenic organisms. Identification of the isolates revealed that the majority of them were belonging to Streptomycetes followed by Nocardia, Micromonospora, Pseudonocardia, Streptosporangium, Nocardiopsis and Saccharomonospora. Among the 33 isolates, Gr1 strain showed antagonistic activity against all checked pathogens. Nine strains showed antibacaterial activity against Listeria, Vibrio cholera, Bacillus cereus, Staphylococcus aureus and Salmonella typhi and only 2 strains (Gr1and Gr25) showed antagonism to E. coli. The overall percentage of activity of actinomycetes isolates against each pathogenic bacterium was also calculated. While 63.63% of the actinomycetes were antagoinistic against Listeria, Vibrio cholerae, and Bacillus cereus, 60.6% of them were antagonistic to Staphylococcus aureus. Very few isolates (6.06%) showed antibacterial activity against E. coli. In general most of the actinomycetes isolates were antagonistic to grampositive bacteria such as Listeria, Bacillus and Staphylococcus than Gram-negative bacteria Vibrio cholerae, E. coli and Salmonella

Optimum growth requirements of nitrifying consortia developed from treated sewage

The optimum growth requirements of two nitrifying consortia developed from treated sewage by enrichment technique were determined by a series of experiments. There was total inhibition of nitrification at above 2.75 g r' NH/- Nand 2.5g r' N02' - N and the ammonia oxidizing consortium preferred a pH at 8.5 and the nitrite oxidizing consortium a pH of 7.5 as the optima for nitrification. Optimum temperatures were between 20° and 30°C for both the groups. As the rate of airnow was increased from I to 7 Ilmin, the build-up of N02- -N increased 10-fold and the consumption of N02- -N increased by a factor of 28.8 implying that the ammonia oxidizing consortium in a bioreactor required three times more aeration than that for nitrite oxidizers for ex pressing their full nitrifying potential. These data directly contribute for developing a fermentati on process for the mass production of nitrifiers as well as for designing bio reactors for nitrifying sewage.

Unconstrained Handwritten Malayalam Character Recognition using Wavelet Transform and Support vector Machine Classifier

This paper presents the application of wavelet processing in the domain of handwritten character recognition. To attain high recognition rate, robust feature extractors and powerful classifiers that are invariant to degree of variability of human writing are needed. The proposed scheme consists of two stages: a feature extraction stage, which is based on Haar wavelet transform and a classification stage that uses support vector machine classifier. Experimental results show that the proposed method is effective

Using Neural Network Classifier Support Vector Machine Regression for the prediction of Melting Point of Drug – like compounds

In our study we use a kernel based classification technique, Support Vector Machine Regression for predicting the Melting Point of Drug – like compounds in terms of Topological Descriptors, Topological Charge Indices, Connectivity Indices and 2D Auto Correlations. The Machine Learning model was designed, trained and tested using a dataset of 100 compounds and it was found that an SVMReg model with RBF Kernel could predict the Melting Point with a mean absolute error 15.5854 and Root Mean Squared Error 19.7576