Qualification: 
Ph.D
Email: 
manjubr@am.amrita.edu

Dr. Manju B. R. currently serves as Associate Professor in the Department of Mathematics, School of Arts & Sciences, Amritapuri. She received the “RIACT -Best Paper Award” and secured 4th Rank in MSc Degree.

Publications

Publication Type: Journal Article

Year of Publication Title

2020

Manju B. R. and R., K. K., “Implementation of Random Projection Filter And Decision Tree J48 For Lung Cancer Detection”, ARPN Journal of Engineering and Applied Sciences, vol. 15, no. 3, 2020.[Abstract]


One of the challenging tasks in this era is the early detection of cancer. The early detection helps to cure the disease completely. Random Projection (RP) is extensively used to reduce the high dimensional features to low dimensional features by projecting data onto a lower space while conserving most of the variation available in the data. J48 can handle both continuous and categorical features and is able to reduce misclassification errors. In this paper we have suggested a method for cancer prediction with the help of different data mining algorithms. The aim is to find out the best filter-classifier combination for the diagnosis. The competency of the algorithms can enhance the insight in to the problem and can thereby minimise the difficulty level in diagnosis.

More »»

2020

Manju B. R. and S., A. V., “Comparative Study Of Datamining Algorithms For Diagnostic Mammograms Using Principal Component Analysis And J48”, ARPN Journal of Engineering and Applied Sciences, vol. 15, no. 3, 2020.[Abstract]


Death rate among women can be considerably brought down with regard to breast cancer if an early detection is viable. The prediction or detection of breast cancer in early stages is a complicated research problem. Using datamining techniques, it is not a difficult task to make it practical. The modern researches show that in most situations these techniques work better than common diagnostic methods. The basic aim of this work is to construct a data demonstrative model which can be used to: predict breast cancer survival even in the presence of missing values in the dataset that can reveal favorable information about the essential factors that determines the chances of survival, and also partition the patients with respect to their common peculiarities. Moreover, to find out a suitable filter-classifier combination. The Principal Component Analysis (PCA) and Decision Tree (J48) are chosen as filters. Further classification process is carried out on filtered dataset using the algorithms Logistic Model Tree (LMT), Random forest and Hoeffding Tree. Decision Tree (J48), were applied to choose the most efficient one. While implementing the classifiers, the dataset for which the feature selection is carried out using PCA gives better classification accuracies. The data mining tool WEKA provides a better platform for required experimental studies. A suitable filter - classifier pair is purposed for breast cancer prognosis by analyzing the results.

More »»

2020

Manju B. R. and Rema, P., “A Performance based comparative study on the Modified version of Empirical Mode Decomposition with traditional Empirical Mode Decomposition”, Procedia Computer Science, vol. 171, pp. 2469-2475, 2020.[Abstract]


Electrocardiogram, popularly known as ECG is a diagnostic test which detects the heart’s electrical system. It is an inevitable diagnostic technique in the medical field which helps to detect any heart related diseases accurately. But there are high chances of addition of any random unwanted signals or noises during the recording process. This hinders the exact diagnosis and hence it is important to denoise these signals. There exist various denoising techniques including Empirical mode decomposition (EMD), Non local means (NLM), various filters, etc. A modified method of EMD is presented here which performs smaller number of iterations and forms only a few Intrinsic mode functions (IMF) when compared with traditional EMD method. The presented work is a performance comparison study between the modified version of EMD which has lesser latency with the traditional EMD.

More »»

2020

Manju B. R. and M.R., S., “ECG Denoising Using Wiener Filter and Kalman Filter”, Third International Conference on Computing and Network Communications (CoCoNet'19), vol. 171, pp. 273 - 281, 2020.[Abstract]


Electrocardiogram (ECG) is a technique of understanding the functioning of heart. Each segment of the ECG signal is significant for the detection of different heart problems. However, some noises generally corrupt the ECG signal. We have performed a research on filters that denoise many kinds of noise observed in real ECG signal. Two filters are implemented to remove the noises, such as Wiener filter and Kalman filter. For better clarity, some performance parameters such as Mean Square Error (MSE), Percentage Root Mean Square Difference (PRD), Signal to Noise Ratio (SNR), Power Spectral Density (PSD), Spectrogram, Magnitude spectrum are used to compare the simulation outcomes. The outcomes of the simulation show that Wiener filter is an outstanding filter for denoising the ECG signal.

More »»

2019

S. Kumar Kashyap, Jain, S., Satheesh, A., Manju B. R., and Jain, V. Kumar, “Controlling the growth rate of cancer cell.”, J Exp Ther Oncol, vol. 13, no. 1, pp. 77-78, 2019.[Abstract]


Objective: To control the rate of growth of the cancer cell is the objective of this paper. In cancer, the rate of the growth of the cancer cell is indefinite. This paper proposes a method to transform into definite rate of growth of the cancer cell from indefinite. This indefiniteness lies with the set of unknown elements. This paper finds these unknown elements by matrix method.

More »»

2016

Manju B. R., “Parameterization of Wavelet Basis in l2(Zn) ”, International Journal of Pharmacy & Technology, vol. 8, no. 4, pp. 22440-22452 , 2016.[Abstract]


The present paper uses the parameterized wavelet basis construction in a very simple representation. The mathematical description in terms of this representation in a single parameter for wavelets of finite length filters helps in modelling out the features of any signal. The explicit formulae for parameterization can be found in [6] The main objective of the paper is to given a varied yet an elegant way to present the wavelet bases. This has been carried out for filter length 4, and hence is representable in a single parameter. Though there are other works on the trigonometric representation of filter coefficients [ 3,4,5,7 ], the formulae presented here forms a convenient way to generate a family of wavelets.

More »»

2012

Manju B. R., Rajan, A. R., and Sugumaran, V., “Optimizing the parameters of wavelets for pattern matching using GA”, Journal of Advanced Research in Engineering and Technology’(IJRET), vol. 3, pp. 75–85, 2012.

2011

Manju B. R., Rajan, A. R., and Sugumaran, V., “New wavelet feature for fault diagnosis of roller bearings using decision tree”, Journal of Mechanical Engineering and Technology, vol. 2, pp. 70–84, 2011.

2010

Manju B. R., Rajan, A. R., and Sugumaran, V., “Wavelet design for fault diagnosis of roller bearings using continuous wavelet transforms”, International Journal of Mechanical Engineering & Technology (IJMET), vol. 1, no. 1, pp. 38–48, 2010.[Abstract]


Fault diagnosis of the roller bearings plays an important role in machine condition monitoring. Researchers are trying to perform the fault diagnosis using machine learning approach. However, there are alternate methods of doing the same. This paper presents the methodology for designing a new wavelet using continuous wavelet transforms. The time domain signals of good bearing is alone required for this study. For the sake of study purpose, three different faults were simulated on to the bearings and the corresponding time domain signals were acquired. The newly designed wavelet gives a classification accuracy of 83.25% in classifying bearing conditions.

More »»

Publication Type: Book Chapter

Year of Publication Title

2020

A. Vinod and Manju B. R., “Optimized Prediction Model to Diagnose Breast Cancer Risk and Its Management”, in Inventive Communication and Computational Technologies, vol. 89, Springer, 2020, pp. 503–515.[Abstract]


Breast malignancy is the second biggest disease that results in fatal condition for women population. Research endeavors have revealed with expanding affirmation that the support vector machines (SVMs) have more noteworthy precise conclusion capacity. In this paper, breast disease determination is dependent on a SVM-based technique that has been proposed. Investigations have been directed on various preparing test allotments of the Wisconsin breast malignancy dataset (WBCD), which is generally utilized among scientists who use machine learning strategies for breast disease conclusion. The working of the technique is assessed by utilizing characterization precision, particularity positive and negative prescient qualities, collector working trademark bends, and perplexity lattice. The outcomes demonstrate that the most elevated grouping precision (99%) is achieved for the SVM.

More »»

Publication Type: Conference Paper

Year of Publication Title

2020

P. M. Warrier, Manju B. R., and Sreedharan, R. P., “A Survey of Pre-processing Techniques Using Wavelets and Empirical-Mode Decomposition on Biomedical Signals”, in Inventive Communication and Computational Technologies, Singapore, 2020.[Abstract]


Recorded biomedical statistics are utilized for predicting various syndromes in humans. Recorded electrical activity of heart can be used for predicting cardiovascular ailment likelihood. Several steps are involved to process biomedical signals, among which the first step related to pre-processing, in which a noisy signal is processed for generating noise-free signal, which can be utilized for further operations. This work gives a detailed understanding of de-noising techniques those have been used for the last decade, for cardiac signals. These techniques utilize the benefits of discrete wavelet transforms (DWT), Bayesian approach, singular value decomposition (SVD), artificial neural networks (ANN), empirical-mode decomposition (EMD), adaptive filtering, and finite impulse response (FIR) filtering. These techniques have been implemented for de-noising of biosignals, individually as well as combining with other techniques, for better results.

More »»

2019

Manju B. R. and Nair, A. R., “Classification of Cardiac Arrhythmia of 12 Lead ECG Using Combination of SMOTEENN, XGBoost and Machine Learning Algorithms”, in 2019 9th International Symposium on Embedded Computing and System Design (ISED), 2019.[Abstract]


Cardiac Arrhythmia is one of those common diseases leading to severe health problems for patients and even sudden death in some cases. Early detection of arrhythmias has a great role in saving lives which can be achieved by analyzing and classifying ECG signal into one of the cardiac arrhythmia. This study gives a method to classify the arrhythmia patients have into one of ten classes, where one class represents the normal condition and the other classes represent various types of arrhythmia conditions. This dataset has been preprocessed. The dataset being highly unbalanced, a combination of oversampling and under sampling using SMOTEENN is applied and feature reduction is carried out using XGboost. The feature reduced dataset is then classified using different supervised learning algorithms of machine learning and an accuracy of 97.48% has occurred which is better than state of art method. This study can be further elaborated using real time data for classification.

More »»
Faculty Research Interest: