Faculty Publications
Permanent URI for this communityhttps://idr.nitk.ac.in/handle/123456789/18736
Publications by NITK Faculty
Browse
29 results
Search Results
Item Accelerating MCMC using model reduction for the estimation of boundary properties within Bayesian framework(Pleiades journals, 2019) Gnanasekaran, N.; Kumar, M.K.In this work, Artificial Neural Network (ANN) and Approximation Error Model (AEM) are proposed as model reduction methods for the simultaneous estimation of the convective heat transfer coefficient and the heat flux from a mild steel fin subject to natural convection heat transfer. The complete model comprises of a three-dimensional conjugate heat transfer from fin whereas the reduced model is simplified to a pure conduction model. On the other hand, the complete model is then replaced with ANN model that acts as a fast forward model. The modeling error that arises due to reduced model is statistically compensated using Approximation Error Model. The estimation of the unknown parameters is then accomplished using the Bayesian framework with Gaussian prior. The sampling space for both the parameters is successfully explored based on Markov chain Monte Carlo method. In addition, the convergence of the Markov chain is ensured using Metropolis–Hastings algorithm. Simulated measurements are used to demonstrate the proposed concept for proving the robustness; finally, the measured temperatures based on in-house experimental setup are then used in the inverse estimation of the heat flux and the heat transfer coefficient for the purpose of validation. © Springer Nature Singapore Pte Ltd. 2019.Item Beyond the data range approach to soft compute the reflection coefficient for emerged perforated semicircular breakwater(Springer, 2019) Kundapura, S.; Hegde, A.V.; Wazerkar, A.V.Prediction of reflection coefficient (Kr) for emerged perforated semicircular breakwater (EPSBW) using artificial neural network (ANN) and adaptive neuro-fuzzy inference systems (ANFIS) is carried out in the present paper. A new approach has been adopted in the present work using ANN and ANFIS models for the prediction of the reflection coefficient (Kr) for the wave periods beyond the range of the dataset used for training the network. The experimental data obtained for a scaled down EPSBW model from regular wave flume experiments at Marine Structure laboratory of National Institute of Technology Karnataka, Surathkal, Mangaluru, India was used. The ensemble was segregated such that certain higher ranges of wave periods were excluded in the training, and possibility of prediction was checked. The independent input parameters (Hi, T, S, D, R, d, hs) that influence the reflection coefficient (Kr) are considered for training as well as testing, where Hi is the incident wave height, T is the wave period, S is the spacing of perforations, D is the diameter of the perforations, R is the radius of the breakwater, d is the depth of the water and hs is the structure height. The accuracy of predictions of reflection coefficient (Kr) is done based on the coefficient of determination (R2), root mean square error (RMSE), and mean absolute error (MAE). The study shows that ANN and ANFIS models may be used for prediction of reflection coefficient Kr of semicircular breakwater for beyond the data range of wave periods used for training. However, ANFIS outperformed ANN model in the prediction of Kr in the case of beyond the data range segregation method. © Springer Nature Singapore Pte Ltd. 2019.Item Stability Analysis of Emerged Seaside Perforated Quarter Circle Breakwater Using Soft Computing Techniques(Springer Science and Business Media Deutschland GmbH, 2022) Madhusoodhanan, S.; Rao, S.Breakwaters are constructed to address a variety of coastal requirements ranging from maintaining tranquility conditions for a port or harbor area to prevent coastal recession. Quarter circle breakwater (QCB) is a composite structure, with a rubble mound foundation and a super structure comprising of a quarter circular surface facing incident waves resting on a horizontal bottom with a rear vertical wall alongside. Be it any structure, it is essential that the design is economic, safe, and functional. Thus, the accurate estimation of minimum (critical) weight of the super structure required to oppose the sliding is vital. Also, physical model studies can be laborious and time-consuming, whereas numerical modeling can be complex. Therefore, under such circumstances, soft computing techniques prove to be handy if sufficient data are available. In this study, W/γHi 2 of an emerged seaside perforated QCB for varying S/D ratios is estimated using ANN, SVM, and AdaBoost models. Hi/gT2, d/hs, and p (%) are chosen as input parameters with the W/γHi 2 as the output parameter. Further, the obtained results are compared using performance indicators such as RMSE, R2, and MAE following which the best model is selected. The data that are used for the present study is collected from the laboratory investigation conducted in the Wave Mechanics Laboratory of the Department of Water Resources and Ocean Engineering, National Institute of Technology Karnataka, Surathkal. © 2022, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.Item Enhancing Cybersecurity: Malicious Webpage Detection Using Machine and Deep Learning(Springer Science and Business Media Deutschland GmbH, 2025) Madhusudhan, R.; Surashe, S.V.; Pravisha, P.A wide range of techniques have been proposed for detecting malicious webpages; however, with the advent of more sophisticated webpage creation processes, it has become more challenging for these approaches to deliver satisfactory outcomes. Blacklisting and classification techniques were used in the past to identify malicious webpages. The classification of the websites becomes more challenging if they are not included on the blacklist. Machine learning techniques are gaining popularity in cybersecurity. One disadvantage of the machine learning model is that it becomes slower when using content-based features. While getting the whois feature, which gives creation, updation, and expiration dates of the webpage, the webpage is physically visited. Hence, there is a chance of malicious activity. Therefore, the process of feature extraction becomes challenging and time-consuming. This article uses the Term Frequency-Inverse Document Frequency (TF-IDF) and Natural Language Processing (NLP) methods to obtain the corpus for benign and malicious words present in the Unified Resource Locator (URL). An artificial neural network (ANN) has been employed to categorize websites as benign or malicious. A comparative analysis of artificial neural networks (ANN) with other machine learning approaches has been conducted. The experimental results demonstrate that ANN has the highest accuracy of 96.70%. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2025.Item Conventional prediction vs beyond data range prediction of loss coefficient for quarter circle breakwater using ANFIS(Springer Verlag service@springer.de, 2015) Hegde, A.V.; Raju, B.Protecting the lagoon area from the wave attack is one of the primary challenges in coastal engineering. Due to the scarcity of rubble and also to achieve economy, new types of breakwaters are being used in place of conventional rubble mound breakwaters. Emerged Perforated Quarter Circle Breakwaters (EPQCB) are artificial concrete breakwaters consisting of a curved perforated face fronting the waves with a vertical wall on rear side and a base slab resting on a low rubble mound base. The perforated curved front face has advantages like energy dissipation and good stability with less material as it is hollow inside. The estimation of hydrodynamic performance characteristics of EPQCB by physical model studies is complex, expensive and time consuming. Hence, computational intelligence (CI) methods are adopted for the evaluation of the performance characteristics like reflection, dissipation, transmission, runup, rundown etc. A number of CI methods like Artificial Neural Network (ANN), Fuzzy logic, and hybrids such as ANFIS, ANN-PCO (particle swarm optimization), ANN-ACO etc., are available and are being used. The paper presents the work carried out to predict the dependent output variable of loss coefficient (Kl) beyond the range of values of one of the input variables i.e., wave period (T) adopted in present work, using the input data on variables of wave height (H), wave period (T), structure height (hs), water depth (d), radius of the breakwater (R), spacing of perforations (S) and diameter of perforations (D) using ANFIS. For this purpose, both the conventional method of data segregation and also a new method called ‘beyond data range’ method are used for both training the ANFIS models and also to predict the dependent variable. Further, the input data was fed to the models in both dimensional and nondimensional form in order to understand the effect of using non-dimensional data in place of dimensional parametric data. The performance of ANFIS models for all the four cases mentioned above was studied and it was found that prediction using conventional method with non-dimensional parameters performed better than other three methods. ANFIS models can be used to predict the performance characteristic Kl of EPQCB beyond the input data range of wave period T. © Springer International Publishing Switzerland 2015.Item Weekly prediction of tides using neural networks(Elsevier Ltd, 2015) Salim, A.M.; Dwarakish, G.S.; Liju, K.V.; Thomas, J.; Devi, G.; Rajeesh, R.Knowledge of tide level is essential for explorations, safe navigation of ships in harbour, disposal of sediments and its movements, environmental observations and in many more coastal engineering applications. Artificial Neural Network (ANN) is being widely applied in coastal engineering field for solving various problems. Its ability to learn highly complex interrelationships based on the provided data sets, along with less amount of data requirement, makes it a powerful modelling tool. The present work is related to predicting the hourly tide levels at Mangalore, Karnataka, using a week's hourly tidal levels as input. The data has been obtained from NMPT, Mangalore and is made use of in predicting tide level using Feed Forward Back Propagation (FFBP) and Non-linear Auto Regressive with eXogenous input (NARX) network. FFBP network yielded correlation coefficient value of 0.564 and NARX network yielded very high correlation coefficient of the order 0.915 for predictions of yearlong hourly tide levels. The study proves that ANN technique can be successfully utilized for the prediction of tides at Mangalore. © 2015 Published by Elsevier Ltd.Item Computational intelligence on hydrodynamic performance characteristics of Emerged Perforated Quarter Circle Breakwater(Elsevier Ltd, 2015) Raju, B.; Hegde, A.V.; Sekhar, O.Protecting the lagoon area from the wave attack is one of the primary challenges in coastal engineering. Due to the scarcity of rubble and to achieve economy, new types of breakwaters are being used in place of conventional rubble mound breakwaters. Emerged Perforated Quarter Circle Breakwater (EPQCB) is an artificial concrete breakwater consisting of a curved perforated face fronting the waves, a vertical wall on back and a base slab resting on a low rubble mound base. The perforated curved front face is having advantages like energy dissipation and good stability with less material as it is hollow inside. Computational Intelligence (CI) can be adopted for the evaluation of performance characteristics like reflection, dissipation, run-up and rundown which are complex, time consuming and expensive to perform in laboratory. The paper presents the work carried out to predict the reflection coefficient (Kr) for input parameters, wave period (T) beyond the data range used for training and of wave height (H) along with the data on input parameters of water depth (d), spacing-perforation ratio (S/D) and radius (R) of the EPQCB. The data on various parameters are taken in two categories for training and testing of ANN as mentioned below in order to understand the effect of using non-dimensional data in place of parametric values: 1) Input in the form of parametric data (H, T, d, R, S, D), and 2) Input in the form of non-dimensional values (H/gT2, d/gT2, S/D, R/H). Better correlation was found when individual dimensional parametric data was used instead of non-dimensional group values in both the methods of prediction. Similarly, the correlation between the beyond the data range prediction and actual values was found to be good in both methods of prediction. © 2015 The Authors. Published by Elsevier Ltd.Item Recognition of repetition and prolongation in stuttered speech using ANN(Springer Science and Business Media Deutschland GmbH info@springer-sbm.com, 2016) Savin, P.S.; Ramteke, P.B.; Koolagudi, S.G.This paper mainly focuses on repetition and prolongation detection in stuttered speech signal. The acoustic and pitch related features like Mel-frequency cepstral coefficients (MFCCs), formants, pitch, zero crossing rate (ZCR) and Energy are used to test the effectiveness in recognizing repetitions and prolongations in stammered speech. Artificial Neural Networks (ANN) are used as classifier. The results are evaluated using combination of different features. The results show that the ANN classifier trained using MFCC features achieves an average accuracy of 87.39% for repetition and prolongation recognition. © Springer India 2016.Item MCMC and approximation error model for the simultaneous estimation of heat flux and heat transfer coefficient using heat transfer experiments(Begell House Inc., 2018) Gnanasekaran, N.; Kumar, M.K.; Balaji, C.This work deals with the simultaneous estimation of the heat flux and the heat transfer coefficient from a mild steel fin losing heat to the ambient by natural convection. Steady state heat transfer experiments are performed on a mild steel fin of dimension 150x250x6 (all dimensions are in mm) placed on to an aluminum base plate of dimension 150x250x8 (all dimensions are in mm). The experimental set up is placed inside a large enclosure to avoid natural disturbances. Nine calibrated K-type thermocouples are used to measure the temperatures of the fin and the base plate. The forward solution of a three dimensional conjugate heat transfer fin model is solved using commercially available ANSYS software in order to obtain the temperature distribution of the fin. An inverse problem is proposed for the estimation of unknown parameters within the Bayesian framework of statistics. Furthermore, a model reduction in the form of Approximation Error Model (AEM) is considered for the inverse conjugate natural convection heat transfer problem. Such an approach not only mitigates the complexity of the inverse problem but also compensates the model reduction with all necessary statistical parameters. Additionally, the sample space within the Bayesian framework is explored with the help of Markov Chain Monte Carlo Method (MCMC) along with the Metropolis-Hastings algorithm. The results of the inverse estimation using Approximation Error Model based on the experimental temperature prove to be a promising alternative in inverse conjugate heat transfer problems. © 2018 International Heat Transfer Conference. All rights reserved.Item Stock price movements classification using machine and deep learning techniques-the case study of indian stock market(Springer Verlag service@springer.de, 2019) Naik, N.; Mohan, B.R.Stock price movements forecasting is an important topic for traders and stock analyst. Timely prediction in stock yields can get more profits and returns. The predicting stock price movement on a daily basis is a difficult task due to more ups and down in the financial market. Therefore, there is a need for a more powerful predictive model to predict the stock prices. Most of the existing work is based on machine learning techniques and considered very few technical indicators to predict the stock prices. In this paper, we have extracted 33 technical indicators based on daily stock price such as open, high, low and close price. This paper addresses the two problems, first is the technical indicator feature selection and identification of the relevant technical indicators by using Boruta feature selection technique. The second is an accurate prediction model for stock price movements. To predict stock price movements we have proposed machine learning techniques and deep learning based model. The performance of the deep learning model is better than the machine learning techniques. The experimental results are significant improves the classification accuracy rate by 5% to 6%. National Stock Exchange, India (NSE) stocks are considered for the experiment. © Springer Nature Switzerland AG 2019.
- «
- 1 (current)
- 2
- 3
- »