Bipin Gogoi - Academia.edu (original) (raw)

Papers by Bipin Gogoi

This thesis presents the detection of time series outliers. The data set used in this work is pro... more This thesis presents the detection of time series outliers. The data set used in this work is provided by the GasDay Project at Marquette University, which produces mathematical models to predict the consumption of natural gas for Local Distribution Companies (LDCs). Flow with no outliers is required to develop and train accurate models. GasDay is using statistical approaches motivated by normally distributed samples such as the 3 − σ rule and the 5 − σ rule to aid the experts in detecting outliers in residuals from the models. However, the Jarque-Bera statistical test shows that the residuals from the GasDay models are not normally distributed. We present an explanation of Density Based Spatial Clustering of Applications with Noise (DBSCAN) and how it is used to detect time series outliers. We have introduced a new application for the DBSCAN algorithm by adapting it to detect outliers in natural gas flow. The performance of DBSCAN is compared with GasDay's existing technique. Five data sets from temperature-sensitive operating areas with identified outliers and 1000 data sets with synthetic outliers are used in the evaluation process. The 1000 synthetic data sets are prepared using the same empirical distribution as one of the identified data set. This work indicates that DBSCAN has shown some improvement in detecting outliers over GasDays existing technique and merits further exploration.

International journal of health sciences

According to WHO report, “Health is a state of complete physical, mental and social well-being an... more According to WHO report, “Health is a state of complete physical, mental and social well-being and not merely the absence of disease or infinity”,[1]. Consequently further developing health status of everything is a piece of population control program. Besides health status can improve or fall apart dependent on the attitudes and activities taken towards health. In demographic analysis, we come across some interesting problems which involve simultaneous consideration of several predictor variables to study their relationships with a dependent variable. Sometimes we are interested to know how well all the variables taken together explain variation in the dependent variable and also how each predictor variable is related to the dependent variable. In this paper, an endeavor has been made to examine the health status among the respondents of Bodo women of Udalguri District, Assam using "Multiple Classification Analysis, MCA". In fact, MCA is a technique for examining the inte...

Social Science Research Network, 2020

A distribution free test is suggested here considering signs of extreme observations in two sampl... more A distribution free test is suggested here considering signs of extreme observations in two sample location problem by taking two subsamples of sizes c and d from the two samples respectively. The suggested test statistics is a linear rank statistics. Power of this test statistics is studied by using Monte Karlo simulation technique. Results are presented in tables.

Scientiae mathematicae Japonicae, 2015

This paper discussed the two well-known procedures to detect single outlier as well as multiple o... more This paper discussed the two well-known procedures to detect single outlier as well as multiple outliers in linear regression on some considered situations. The procedure is Rn statistics for single outlier and Marasinghe procedure for multiple outliers. We have calculated the power of these test statistics using Monte Carlo simulation method and conclusions are made based on the calculated results.

This paper attempt to estimate the trend values for the production of Rice in Assam (India) for p... more This paper attempt to estimate the trend values for the production of Rice in Assam (India) for period 1990 to 2010. A time series analysis has been made to estimate the trend values for the production of Rice for the period of 19990 to 2010, Assam, India. The data have been collected from the official record of the office of the Extension coordinator Agricultural Technology Management Agency (ATMA) Dhemaji District of Assam (India) with personal investigation. The natures of the data are secondary.

Usually when we have to deal with an analysis of a particular data set, we often choose a statist... more Usually when we have to deal with an analysis of a particular data set, we often choose a statistical procedure on the reference of the analysis. Now if in this analysis the parametric assumptions are fulfilled we will prefer F test for the two sample scale problem. But if the behaviour of the data set does not fulfil these assumptions then we have to focus on some other tests. In this paper it is tried to throw light on the use of Adaptive tests on the context of this problem. This paper comprises of power comparison between F test, some traditional NonParametric tests and proposed Adaptive tests. Under the experiment equal and unequal sample sizes are taken considering different alternatives. This is done by the use of Monte Carlo simulation technique. At the end of the experiment, inspite of many limitations, some satisfactory results are obtained. All those results are displayed in tabular and graphical form.

Human Cognition in Evolution & Development eJournal, 2020

The purpose of this article to propose a comparative study in estimation methods under regression... more The purpose of this article to propose a comparative study in estimation methods under regression analysis between adaptive, OLS, M and S estimation.. An example is used here to fulfill the objective. In this example S estimation is more suitable. At the end of the discussion we find that if the parent distribution of the data set is not normal then Adaptive estimation method is more suitable. Results are displayed here in tables.

International Journal of Computational and Theoretical Statistics, 2015

The analysis of data for outliers is a part of model building and data summarizing for model test... more The analysis of data for outliers is a part of model building and data summarizing for model testing, parameter estimation, prediction and peculiarity investigations. Any influential point can disproportionately pull the ordinary least squares line and distort the predictions. Thus the detection of outlying observations is very essential in the course of model building in various disciplines, such as, medical research, economics, sociology, computer science, etc. A point is an influential one if it causes dramatic change in the model after its deletion. Each of the available test statistics has different cutoff values that indicate the amount of outlyingness. Sometimes only one statistic is sufficient to provide the information about influential points but often it is necessary to examine the cutoff of more than one influence measure. The reason behind is that all the cutoff values are either a function of the sample size or number of predictors or both. Also validity of the cutoff value is subjected to some additional conditions. In this paper we try to critically examine those conditions with the help of simulation study. We shall use a few combinations of () , where is the sample size and is the no. of outliers for assessing the performances.

International Journal of Science and Research (IJSR), 2016

One of the more common statistical tests of significance is the test for paired data. Paired data... more One of the more common statistical tests of significance is the test for paired data. Paired data can be obtained from experimental units when measurements are obtained before and after the administration of a treatment. In this paper, comparison is made between adaptive tests, non parametric wilcoxon test and traditional paired sample t test. For comparison of power, Monte-Carlo simulation method is used here. The new adaptive procedure is shown to preserve the size of the test at its nominal level for all continuous distributions. This research work finally comes to the conclusion that in case of normal distribution traditional t test is more powerful than the others. Whereas in case of long tailed or skewed distribution the adaptive tests are more powerful.

This thesis presents the detection of time series outliers. The data set used in this work is pro... more This thesis presents the detection of time series outliers. The data set used in this work is provided by the GasDay Project at Marquette University, which produces mathematical models to predict the consumption of natural gas for Local Distribution Companies (LDCs). Flow with no outliers is required to develop and train accurate models. GasDay is using statistical approaches motivated by normally distributed samples such as the 3 − σ rule and the 5 − σ rule to aid the experts in detecting outliers in residuals from the models. However, the Jarque-Bera statistical test shows that the residuals from the GasDay models are not normally distributed. We present an explanation of Density Based Spatial Clustering of Applications with Noise (DBSCAN) and how it is used to detect time series outliers. We have introduced a new application for the DBSCAN algorithm by adapting it to detect outliers in natural gas flow. The performance of DBSCAN is compared with GasDay's existing technique. Five data sets from temperature-sensitive operating areas with identified outliers and 1000 data sets with synthetic outliers are used in the evaluation process. The 1000 synthetic data sets are prepared using the same empirical distribution as one of the identified data set. This work indicates that DBSCAN has shown some improvement in detecting outliers over GasDays existing technique and merits further exploration.

The present paper considers power comparison between two sample Adaptive tests and traditional no... more The present paper considers power comparison between two sample Adaptive tests and traditional non parametric tests in the field of location problem. In the proposed approach it is shown by MonteCarlo simulation study, that the Adaptive tests are superior to the other existing competitors in terms of both robustness of size and power.

In this paper we try to study the power of some tests statistics viz. F-Test, Hettmansperger test... more In this paper we try to study the power of some tests statistics viz. F-Test, Hettmansperger test based on rank and Normal Score test ,Test based on long tailed and test based on short tailed score. We have used Monte-Carlo simulation technique for our study. We calculated empirical level and power of the tests for different combination of parameters generating observation from various distributions. Based on the obtained results discussion are made and conclusion are given accordingly.

The purpose of this article to propose a comparative study in estimation methods for location par... more The purpose of this article to propose a comparative study in estimation methods for location parameter between adaptive and some other methods. Underlying distributions for this study are Normal, exponential and logistic. Adaptive estimation methods are found to give more accurate results for asymmetric distributions compared to symmetric distributions. Monte Carlo simulation method is used here to fulfill the objectives. Results are displayed here in tables.

International Journal of Computational and Theoretical Statistics

The analysis of data from block designs arises in different fields. Various parametric and nonpar... more The analysis of data from block designs arises in different fields. Various parametric and nonparametric procedures are available for analysis of such data. But analysis of such design with censored data has not been as widely studied.. In this paper we want to study the performance of few existing tests when some of the observations are right censored. Two different kinds of scores are used for this purpose. Results are obtained through simulation Discussion are made based on the obtained results. For few situations results are shown graphically for easy visual comparison.

IOSR Journal of Mathematics

In the presence of outliers least square estimation is insufficient and can be biased. Adaptive e... more In the presence of outliers least square estimation is insufficient and can be biased. Adaptive estimation is robust estimation that can be used to improve the accuracy of the estimate by reducing the influence of outliers. Adaptive estimators can be effective in achieving low mean squared error for a variety of non normal distributions of errors. M estimation is the extension of the maximum likelihood estimation and is also a robust estimation. In this paper, comparison is made between OLS Estimation, Adaptive estimation and M estimation under a particular situation (An example).

IARJSET, 2016

For the two samples scale problem, we have considered the ten existing tests viz. F-test, Mood te... more For the two samples scale problem, we have considered the ten existing tests viz. F-test, Mood test, Ansari-Bradley test, Siegel-Tukey test, Lepage type test, Cramer-Von-Mises test, Kamat test, Wald-W test, Wald-R test and Levene test. To study the performance of these tests we generate data from the distributions viz. Normal, Exponential, Double Exponential, Logistic and Cauchy. We have calculated the empirical level under null situations and power under alternative situations taking various combinations of scale parameters and sample sizes using simulation technique. We also simulated critical values for Kamat test, Wald-W test and Wald-R test for some sample sizes. Results are displayed in various tables and graphs. Discussion and conclusion are made on the basis of results obtained.

IARJSET, 2016

Performance of Mack-Wolfe type test using different scores suitable for long tail, short tail, ri... more Performance of Mack-Wolfe type test using different scores suitable for long tail, short tail, right skewed distributions under Umbrella alternatives are studied. We also compare the performance of this type test with the Jonckheere test of ordered alternative and general classical F-statistic considering various combinations of parameters. Results are obtained by simulation technique repeating 10,000 times. These results are presented in different tables. Results are also shown graphically for easy visual comparison.

IOSR Journal of Mathematics

The uniform distribution appears due to natural random events or to the application of methods fo... more The uniform distribution appears due to natural random events or to the application of methods for transforming samples from any other distribution to samples with uniformly distributed values in the interval (0,1). Thus, in order to test whether a samplecomes from a given distribution, one can test whether its transformed sample is distributed according to the uniform distribution or not. Several test procedures are developed to test the goodness of fit for uniformity. In this paper, we want to study the performance of eleven different tests for uniformity by considering different sample sizes as well as different alternatives. The results so obtained are displayed in various tables and graphs. Finally, conclusions are made on the basis of the results.

This thesis presents the detection of time series outliers. The data set used in this work is pro... more This thesis presents the detection of time series outliers. The data set used in this work is provided by the GasDay Project at Marquette University, which produces mathematical models to predict the consumption of natural gas for Local Distribution Companies (LDCs). Flow with no outliers is required to develop and train accurate models. GasDay is using statistical approaches motivated by normally distributed samples such as the 3 − σ rule and the 5 − σ rule to aid the experts in detecting outliers in residuals from the models. However, the Jarque-Bera statistical test shows that the residuals from the GasDay models are not normally distributed. We present an explanation of Density Based Spatial Clustering of Applications with Noise (DBSCAN) and how it is used to detect time series outliers. We have introduced a new application for the DBSCAN algorithm by adapting it to detect outliers in natural gas flow. The performance of DBSCAN is compared with GasDay's existing technique. Five data sets from temperature-sensitive operating areas with identified outliers and 1000 data sets with synthetic outliers are used in the evaluation process. The 1000 synthetic data sets are prepared using the same empirical distribution as one of the identified data set. This work indicates that DBSCAN has shown some improvement in detecting outliers over GasDays existing technique and merits further exploration.

International journal of health sciences

According to WHO report, “Health is a state of complete physical, mental and social well-being an... more According to WHO report, “Health is a state of complete physical, mental and social well-being and not merely the absence of disease or infinity”,[1]. Consequently further developing health status of everything is a piece of population control program. Besides health status can improve or fall apart dependent on the attitudes and activities taken towards health. In demographic analysis, we come across some interesting problems which involve simultaneous consideration of several predictor variables to study their relationships with a dependent variable. Sometimes we are interested to know how well all the variables taken together explain variation in the dependent variable and also how each predictor variable is related to the dependent variable. In this paper, an endeavor has been made to examine the health status among the respondents of Bodo women of Udalguri District, Assam using "Multiple Classification Analysis, MCA". In fact, MCA is a technique for examining the inte...

Social Science Research Network, 2020

A distribution free test is suggested here considering signs of extreme observations in two sampl... more A distribution free test is suggested here considering signs of extreme observations in two sample location problem by taking two subsamples of sizes c and d from the two samples respectively. The suggested test statistics is a linear rank statistics. Power of this test statistics is studied by using Monte Karlo simulation technique. Results are presented in tables.

Scientiae mathematicae Japonicae, 2015

This paper discussed the two well-known procedures to detect single outlier as well as multiple o... more This paper discussed the two well-known procedures to detect single outlier as well as multiple outliers in linear regression on some considered situations. The procedure is Rn statistics for single outlier and Marasinghe procedure for multiple outliers. We have calculated the power of these test statistics using Monte Carlo simulation method and conclusions are made based on the calculated results.

This paper attempt to estimate the trend values for the production of Rice in Assam (India) for p... more This paper attempt to estimate the trend values for the production of Rice in Assam (India) for period 1990 to 2010. A time series analysis has been made to estimate the trend values for the production of Rice for the period of 19990 to 2010, Assam, India. The data have been collected from the official record of the office of the Extension coordinator Agricultural Technology Management Agency (ATMA) Dhemaji District of Assam (India) with personal investigation. The natures of the data are secondary.

Usually when we have to deal with an analysis of a particular data set, we often choose a statist... more Usually when we have to deal with an analysis of a particular data set, we often choose a statistical procedure on the reference of the analysis. Now if in this analysis the parametric assumptions are fulfilled we will prefer F test for the two sample scale problem. But if the behaviour of the data set does not fulfil these assumptions then we have to focus on some other tests. In this paper it is tried to throw light on the use of Adaptive tests on the context of this problem. This paper comprises of power comparison between F test, some traditional NonParametric tests and proposed Adaptive tests. Under the experiment equal and unequal sample sizes are taken considering different alternatives. This is done by the use of Monte Carlo simulation technique. At the end of the experiment, inspite of many limitations, some satisfactory results are obtained. All those results are displayed in tabular and graphical form.

Human Cognition in Evolution & Development eJournal, 2020

The purpose of this article to propose a comparative study in estimation methods under regression... more The purpose of this article to propose a comparative study in estimation methods under regression analysis between adaptive, OLS, M and S estimation.. An example is used here to fulfill the objective. In this example S estimation is more suitable. At the end of the discussion we find that if the parent distribution of the data set is not normal then Adaptive estimation method is more suitable. Results are displayed here in tables.

International Journal of Computational and Theoretical Statistics, 2015

The analysis of data for outliers is a part of model building and data summarizing for model test... more The analysis of data for outliers is a part of model building and data summarizing for model testing, parameter estimation, prediction and peculiarity investigations. Any influential point can disproportionately pull the ordinary least squares line and distort the predictions. Thus the detection of outlying observations is very essential in the course of model building in various disciplines, such as, medical research, economics, sociology, computer science, etc. A point is an influential one if it causes dramatic change in the model after its deletion. Each of the available test statistics has different cutoff values that indicate the amount of outlyingness. Sometimes only one statistic is sufficient to provide the information about influential points but often it is necessary to examine the cutoff of more than one influence measure. The reason behind is that all the cutoff values are either a function of the sample size or number of predictors or both. Also validity of the cutoff value is subjected to some additional conditions. In this paper we try to critically examine those conditions with the help of simulation study. We shall use a few combinations of () , where is the sample size and is the no. of outliers for assessing the performances.

International Journal of Science and Research (IJSR), 2016

One of the more common statistical tests of significance is the test for paired data. Paired data... more One of the more common statistical tests of significance is the test for paired data. Paired data can be obtained from experimental units when measurements are obtained before and after the administration of a treatment. In this paper, comparison is made between adaptive tests, non parametric wilcoxon test and traditional paired sample t test. For comparison of power, Monte-Carlo simulation method is used here. The new adaptive procedure is shown to preserve the size of the test at its nominal level for all continuous distributions. This research work finally comes to the conclusion that in case of normal distribution traditional t test is more powerful than the others. Whereas in case of long tailed or skewed distribution the adaptive tests are more powerful.

This thesis presents the detection of time series outliers. The data set used in this work is pro... more This thesis presents the detection of time series outliers. The data set used in this work is provided by the GasDay Project at Marquette University, which produces mathematical models to predict the consumption of natural gas for Local Distribution Companies (LDCs). Flow with no outliers is required to develop and train accurate models. GasDay is using statistical approaches motivated by normally distributed samples such as the 3 − σ rule and the 5 − σ rule to aid the experts in detecting outliers in residuals from the models. However, the Jarque-Bera statistical test shows that the residuals from the GasDay models are not normally distributed. We present an explanation of Density Based Spatial Clustering of Applications with Noise (DBSCAN) and how it is used to detect time series outliers. We have introduced a new application for the DBSCAN algorithm by adapting it to detect outliers in natural gas flow. The performance of DBSCAN is compared with GasDay's existing technique. Five data sets from temperature-sensitive operating areas with identified outliers and 1000 data sets with synthetic outliers are used in the evaluation process. The 1000 synthetic data sets are prepared using the same empirical distribution as one of the identified data set. This work indicates that DBSCAN has shown some improvement in detecting outliers over GasDays existing technique and merits further exploration.

The present paper considers power comparison between two sample Adaptive tests and traditional no... more The present paper considers power comparison between two sample Adaptive tests and traditional non parametric tests in the field of location problem. In the proposed approach it is shown by MonteCarlo simulation study, that the Adaptive tests are superior to the other existing competitors in terms of both robustness of size and power.

In this paper we try to study the power of some tests statistics viz. F-Test, Hettmansperger test... more In this paper we try to study the power of some tests statistics viz. F-Test, Hettmansperger test based on rank and Normal Score test ,Test based on long tailed and test based on short tailed score. We have used Monte-Carlo simulation technique for our study. We calculated empirical level and power of the tests for different combination of parameters generating observation from various distributions. Based on the obtained results discussion are made and conclusion are given accordingly.

The purpose of this article to propose a comparative study in estimation methods for location par... more The purpose of this article to propose a comparative study in estimation methods for location parameter between adaptive and some other methods. Underlying distributions for this study are Normal, exponential and logistic. Adaptive estimation methods are found to give more accurate results for asymmetric distributions compared to symmetric distributions. Monte Carlo simulation method is used here to fulfill the objectives. Results are displayed here in tables.

International Journal of Computational and Theoretical Statistics

The analysis of data from block designs arises in different fields. Various parametric and nonpar... more The analysis of data from block designs arises in different fields. Various parametric and nonparametric procedures are available for analysis of such data. But analysis of such design with censored data has not been as widely studied.. In this paper we want to study the performance of few existing tests when some of the observations are right censored. Two different kinds of scores are used for this purpose. Results are obtained through simulation Discussion are made based on the obtained results. For few situations results are shown graphically for easy visual comparison.

IOSR Journal of Mathematics

In the presence of outliers least square estimation is insufficient and can be biased. Adaptive e... more In the presence of outliers least square estimation is insufficient and can be biased. Adaptive estimation is robust estimation that can be used to improve the accuracy of the estimate by reducing the influence of outliers. Adaptive estimators can be effective in achieving low mean squared error for a variety of non normal distributions of errors. M estimation is the extension of the maximum likelihood estimation and is also a robust estimation. In this paper, comparison is made between OLS Estimation, Adaptive estimation and M estimation under a particular situation (An example).

IARJSET, 2016

For the two samples scale problem, we have considered the ten existing tests viz. F-test, Mood te... more For the two samples scale problem, we have considered the ten existing tests viz. F-test, Mood test, Ansari-Bradley test, Siegel-Tukey test, Lepage type test, Cramer-Von-Mises test, Kamat test, Wald-W test, Wald-R test and Levene test. To study the performance of these tests we generate data from the distributions viz. Normal, Exponential, Double Exponential, Logistic and Cauchy. We have calculated the empirical level under null situations and power under alternative situations taking various combinations of scale parameters and sample sizes using simulation technique. We also simulated critical values for Kamat test, Wald-W test and Wald-R test for some sample sizes. Results are displayed in various tables and graphs. Discussion and conclusion are made on the basis of results obtained.

IARJSET, 2016

Performance of Mack-Wolfe type test using different scores suitable for long tail, short tail, ri... more Performance of Mack-Wolfe type test using different scores suitable for long tail, short tail, right skewed distributions under Umbrella alternatives are studied. We also compare the performance of this type test with the Jonckheere test of ordered alternative and general classical F-statistic considering various combinations of parameters. Results are obtained by simulation technique repeating 10,000 times. These results are presented in different tables. Results are also shown graphically for easy visual comparison.

IOSR Journal of Mathematics

The uniform distribution appears due to natural random events or to the application of methods fo... more The uniform distribution appears due to natural random events or to the application of methods for transforming samples from any other distribution to samples with uniformly distributed values in the interval (0,1). Thus, in order to test whether a samplecomes from a given distribution, one can test whether its transformed sample is distributed according to the uniform distribution or not. Several test procedures are developed to test the goodness of fit for uniformity. In this paper, we want to study the performance of eleven different tests for uniformity by considering different sample sizes as well as different alternatives. The results so obtained are displayed in various tables and graphs. Finally, conclusions are made on the basis of the results.