ISI Digital Commons (Indian Statistical Institute )
Not a member yet
7571 research outputs found
Sort by
LIFA: Language identification from audio with LPCC-G features
No full textIn Western countries, speech recognition-based technologies have significantly developed compared to the countries of the South Asian subcontinent like India. India is a multilingual country (22 scheduled languages) with over 1.3 Billion population of which a major percentage faces difficulty with the user interface of different technological advancements and therefore speech recognition tools are very useful. In this paper, we propose LIFA: Language Identification From Audio - a fully automated tool that can identify the spoken language (phrases/words) and invoke the language-specific recognition engine. Experiments were performed on more than 2200 hours of data from the top-11 spoken languages in India. The clips were parameterized with a novel linear predictive cepstral coefficient (LPCC)-based features, which we call LPCC-Grade (LPCC-G). The proposed feature is capable of focusing on the distribution of energy across different frequency ranges in an audio clip for better classification while avoiding high dimensionality issues. Using a random forest-based classifier, we achieved the highest accuracy of 99.01%. Further, we tested the robustness of the system with different noisy scenarios on multiple datasets wherein accuracies in the range of 79%-98% were obtained. We also studied other popular existing features in our comparison where accuracies of 96.37% and 92.48% were obtained for LSF and MFCC-based features
Longest increasing path within the critical strip
No full textA Poisson point process of unit intensity is placed in the square [0, n]2. An increasing path is a curve connecting (0, 0) with (n, n) which is non-decreasing in each coordinate. Its length is the number of points of the Poisson process which it passes through. Baik, Deift and Johansson proved that the maximal length of an increasing path has expectation 2n − n1/3(c1 + o(1)), variance n2/3(c2 + o(1)) for some c1, c2 \u3e 0 and that it converges to the Tracy–Widom distribution after suitable scaling. Johansson further showed that all maximal paths have a displacement of n23+o(1) from the diagonal with probability tending to one as n → ∞. Here we prove that the maximal length of an increasing path restricted to lie within a strip of width nγ, γ\u3c23, around the diagonal has expectation 2n − n1−γ+o(1), variance n1−γ2+o(1) and that it converges to the Gaussian distribution after suitable scaling
Modulation of charge and spin circular currents in a ring-wire hybrid setup
No full textWe present a comprehensive investigation into the charge and spin circular currents in a mesoscopic hybrid system, with a particular focus on the intricate interplay between the Aubry-André-Harper (AAH) potential, Aharonov-Bohm (AB) flux, chemical potential (μ), and antiferromagnetic (AF) ordering. The proposed quantum system comprises a composite structure of an AF ring coupled to an AAH chain. Utilizing a tight-binding model and operator method to calculate charge and spin circular currents, we uncover a range of intriguing phenomena. An interesting finding is that, while the antiferromagnetic ring alone does not exhibit spin channel separation due to the symmetry between the up and down spin sub-Hamiltonians, the introduction of a dangling bond or chain can break this symmetry, leading to a spin separation effect. The AAH potential in the chain disrupts energy level symmetry, significantly impacting transport behavior. The charge circular current exhibits periodic oscillations with the AB flux, and its polarity changes with variations in μ. The AAH phase affects the charge current through the shifting of energy levels. Furthermore, the spin current displays oscillatory behavior as the AAH potential strength changes, with peaks emerging due to interference phenomena caused by disorder-induced localization. A possible experimental realization of our proposed quantum setup is also discussed, for the sake of completeness. This work may provide important insights into the complex physics of charge and spin transport in various hybrid mesoscopic systems, offering promising avenues for future research and technological applications
NEW ACCELERATED MODULUS-BASED ITERATION METHOD FOR SOLVING LARGE AND SPARSE LINEAR COMPLEMENTARITY PROBLEM
No full textFor the large and sparse linear complementarity problem, we provide a family of new accelerated modulus-based iteration methods in this article. We provide some sufficient criteria for the convergence analysis when the system matrix is a P-matrix or an H+-matrix. In addition, we provide some numerical examples of the different parameters to illustrate the efficacy of our proposed methods. These methods help us reduce the number of iterations and the time required by the CPU, which improves convergence performance
Normal approximation for statistics of randomly weighted complexes
No full textWe prove normal approximation bounds for statistics of randomly weighted (simplicial) complexes. In particular, we consider the complete d-dimensional complex on n vertices with d-simplices equipped with i.i.d. weights. Our normal approximation bounds are quantified in terms of stabilization of difference operators, i.e., the effect on the statistic under addition/deletion of simplices. Our proof is based on Chatterjee’s normal approximation bound and is a higher-dimensional analogue of the work of Cao on sparse Erdős–Rényi random graphs but our bounds are more in the spirit of ‘quantitative two-scale stabilization’ bounds by Lachièze-Rey, Peccati, and Yang. As applications, we prove a CLT for nearest face-weights in randomly weighted d-complexes and give a normal approximation bound for local statistics of random d-complexes
Nutritional status of tribal and non-tribal school-going children in rural Bangladesh: A comparative study
No full textBackground: Inadequate nutrition of school-going children is a major concern in Bangladesh, and it can negatively affect their productivity. It is important to consider the food pattern, socio-cultural, and economic differences between tribal (T) and non-tribal (NT) communities in Bangladesh when evaluating their nutritional status. This study aimed to investigate the nutritional status of school-going children in the rural area of Rajshahi district’s High Barind Tract (HBT) region of Bangladesh. Additionally, we compared the nutritional status between T and NT school-going children in the same area. Methods: This was a cross sectional household study. Data were collected from T and NT households in the HBT region in the Rajshahi district of Bangladesh, from January to June of 2019. A total of 500 (T 81, NT 419) school-going children aged 6–13 years were selected as samples using mixed sampling, including convenience sampling (non-probability) and simple random sampling (probability) methods. Nutritional status was assessed using body mass index-for-age z-score (BAZ) and height-for-age z-score (HAZ) according to WHO guidelines. Thinness was defined as BAZ \u3c -2SD and stunting as HAZ \u3c -2SD. Descriptive statistics, Z-proportional test, and logistic regression model were used to analyze the effect of selected independent variables on nutritional status of T and NT children. Results: Among school-going children, 15.20% were suffering from thinness (T 12.30% and NT 15.80%) and 17.80% stunting (T 13.60% and NT 18.60%), respectively. The difference in thinness (p \u3e 0.05) and stunting (p \u3e 0.05) were not significant between T and NT. The distribution of BAZ and HAZ of T and NT children were normally distributed, and were positioned negatively compared to the WHO standards. The logistic model identified the following factors for thinness: (i) mother with non-or-primary education (aOR = 1.89, 95% CI: 1.05–3.43, p \u3c 0.05), (ii) underweight mother (aOR = 3.86, 95% CI: 1.48–10.06, p \u3c 0.01), and (iii) underweight father (aOR = 4.12, 95% CI: 1.50-11.29, p \u3c 0.01). For stunting, the factors were: (i) mother as a housewife (aOR = 2.79, 95% CI: 1.16–6.71, p \u3c 0.05), (ii) father working as labour (aOR = 1.77, 95% CI: 1.01–3.278, p \u3c 0.05), (iii) severe food insecurity in the household (aOR = 2.37, 95% CI: 1.23–4.54, p \u3c 0.05), and (iv) children playing outside regularly more than 2 h (aOR = 2.19, 95% CI: 1.31–3.67, p \u3c 0.01). Conclusion: In rural Bangladesh, the nutritional status of T and NT school-going children did not show significant defferences. However, the mean z-score values for both groups of children were lower than the WHO standard, indicating that both communities have poor nutritional status
On approximation and estimation of distribution function of sum of independent random variables
No full textIn this paper, we obtain an approximation for the distribution function of sum of two independent random variables using quantile based representation. The error of approximation is shown to be negligible under some mild conditions. We then use the approximation to obtain a non-parametric estimator for the distribution function of sum of two independent random variables. The exact distribution of the proposed estimator is derived. The estimator is shown to be strongly consistent and asymptotically normally distributed. Extensive Monte Carlo simulation studies are carried out to evaluate the bias and mean squared error of the estimator and also to assess the approximation error. We also compare the performance of the proposed estimator with other estimators available in the literature. Finally, we illustrate the use of the proposed estimator for estimating the reliability function of a standby redundant system
ON CHARACTERIZATION AND RECOGNITION OF PROPER TAGGED PROBE INTERVAL GRAPHS
No full textInterval graphs were used in the study of the human genome project by the molecular biologist Benzer. Later on probe interval graphs were introduced by Zhang as a generalization of interval graphs for the study of cosmid contig mapping of DNA. Further research in this area required more useful and cost-effective tools. The concept of tagged probe interval graphs is motivated from this point of view. In this paper, we consider a natural subclass of it, namely, the class of proper tagged probe interval graphs. In this paper, we present a characterization theorem and a linear time recognition algorithm for proper tagged probe interval graphs. Also, we discuss the interrelations between the classes of proper tagged probe interval graphs and tagged probe interval graphs with probe interval graphs and probe proper interval graphs
On Completely Mixed Games
No full textA matrix game is considered completely mixed if all the optimal pairs of strategies in the game are completely mixed. In this paper, we establish that a matrix game A, with a value of zero, is completely mixed if and only if the value of the game associated with A+Di is positive for all i, where Di represents a diagonal matrix where ith diagonal entry is 1 and else 0. Additionally, we address Kaplansky’s question from 1945 regarding whether an odd-ordered symmetric game can be completely mixed, and provide characterizations for odd-ordered skew-symmetric matrices to be completely mixed. Moreover, we demonstrate that if A is an almost skew-symmetric matrix and the game associated with A has value positive, then A+Di∈Q for all i, where Di is a diagonal matrix whose ith diagonal entry is 1 and else 0. Skew-symmetric matrices and almost skew-symmetric matrices with value positive fall under the class of P0 and Q0, making them amenable to processing through Lemke’s algorithm
On some non parametric estimators of the quantile density function for a stationary associated process
No full textIn this article, we consider smooth estimators for the quantile density function (qdf) for a sequence (Formula presented.) of stationary non negative associated random variables with a common marginal distribution function. The qdf is given by (Formula presented.), (Formula presented.) representing the corresponding quantile function. The smooth estimators of (Formula presented.) considered here are adapted from those of (Formula presented.) considered in Chaubey, Dewan, and Li (2021). A few asymptotic properties of these estimators are established parallel to those in the i.i.d. case. A numerical study comparing the mean squared errors of various estimators indicates the advantages and a few limitations of various estimators. The smoothing parameter is selected based on the BCV and RLCV (a variation of likelihood cross-validation) criteria. It is concluded, based on the numerical studies, that the RLCV criterion may produce over-smoothing, hence BCV criterion may be preferable. The numerical studies also suggest that, overall, the estimator proposed by Soni, Dewan, and Jain (2012) seems to have some advantage over the other estimators considered in this article