Sumet Mehta

Followers

Following

Co-author

Public Views

Interests

Uploads

Papers by Sumet Mehta

Autoencoder based Adaptive Multi Objective Particle Swarm Optimization for Gene Selection

by IJIRCST I and Sumet Mehta

International Journal of Innovative Research in Computer Science and Technology (IJIRCST), 2025

In gene expression analysis, selecting informative genes is essential for uncovering biological m... more In gene expression analysis, selecting informative genes is essential for uncovering biological mechanisms and identifying potential biomarkers. However, conventional gene selection methods often struggle with scalability and parameter tuning, limiting their effectiveness in large-scale datasets and algorithmic optimization. To overcome these challenges, we propose Autoencoder-based Adaptive Multi-Objective Particle Swarm Optimization for Gene Selection (AAMOPSO). Our approach incorporates an autoencoder-based preprocessing step to enhance scalability by learning a compressed representation of gene expression data, reducing dimensionality while retaining critical features. Additionally, we introduce an Adaptive Parameter Tuning mechanism within the Multi-Objective Particle Swarm Optimization (MOPSO) framework, dynamically adjusting algorithm parameters based on real-time performance metrics. Extensive experiments on four benchmark microarray datasets demonstrate that AAMOPSO consistently outperforms existing state-of-the-art methods in classification accuracy and the compactness of selected gene subsets.

Download

Robust kernel ensemble regression in diversified kernel space with shared parameters

Applied Intelligence

A Comparative Analysis on Handling Big Data Using Cloud Services

Lahore Garrison University Research Journal of Computer Science and Information Technology

In this era of technology, a lot of advancements have been done in almost every field such as med... more In this era of technology, a lot of advancements have been done in almost every field such as medical, science, aerospace and other fields. With the increasing advancements in technology, a lot of data is being produced at the same time. For instances in the field of medicine there is a huge amount of data that is being generated as there are hundreds and thousands of patients who came for their checkup. So now the question arises where this huge amount of data is being stored. This huge amount of data is called as Big Data. And the major problem faced is how to manage and organize this huge amount of data along with its security and not being lost. Big data is used for extracting a lot of useful information but it is not easy to organize it. If the data is being lost than there are a lot of problems that can occur on a huge level as a lot of data being stored in big data is very confidential. This data can be stored on cloud which is the new advancement in the field of technology a...

Download

A Technique and Architectural Design for Criminal Detection based on Lombroso Theory Using Deep Learning

Lahore Garrison University Research Journal of Computer Science and Information Technology, 2020

Crimes and criminal activities are increasing day by day and there are no proper criteria to sear... more Crimes and criminal activities are increasing day by day and there are no proper criteria to search, detect, identify, and predict these criminals. Despite various surveillance cameras in different areas still, crimes are at a peak. The police investigation department cannot efficiently detect the criminals in time. However, in many countries for the sake of public and private security, the initiation of security technologies has been employed for criminal identification or recognition with the help of footprint identification, fingerprint identification, facial recognition, or based on other suspicious activity detections through surveillance cameras. However, there are limited automated systems that can identify the criminals precisely and get the accurate or precise similarity between the recorded footage images with the criminals that already are available in the police criminal records. To make the police investigation department more effective, this research work presents the ...

Download

Generalized Multi-manifold Graph Ensemble Embedding for Multi-View Dimensionality Reduction

Lahore Garrison University Research Journal of Computer Science and Information Technology, 2020

In this paper, we propose a new dimension reduction (DR) algorithm called ensemble graph-based lo... more In this paper, we propose a new dimension reduction (DR) algorithm called ensemble graph-based locality preserving projections (EGLPP); to overcome the neighborhood size k sensitivity in locally preserving projections (LPP). EGLPP constructs a homogeneous ensemble of adjacency graphs by varying neighborhood size k and finally uses the integrated embedded graph to optimize the low-dimensional projections. Furthermore, to appropriately handle the intrinsic geometrical structure of the multi-view data and overcome the dimensionality curse, we propose a generalized multi-manifold graph ensemble embedding framework (MLGEE). MLGEE aims to utilize multi-manifold graphs for the adjacency estimation with automatically weight each manifold to derive the integrated heterogeneous graph. Experimental results on various computer vision databases verify the effectiveness of proposed EGLPP and MLGEE over existing comparative DR methods.

Download

MKEL: Multiple Kernel Ensemble Learning via Unified Ensemble Loss for Image Classification

ACM Transactions on Intelligent Systems and Technology

In this article, a novel ensemble model, called Multiple Kernel Ensemble Learning (MKEL), is deve... more In this article, a novel ensemble model, called Multiple Kernel Ensemble Learning (MKEL), is developed by introducing a unified ensemble loss. Different from the previous multiple kernel learning (MKL) methods, which attempt to seek a linear combination of basis kernels as a unified kernel, our MKEL model aims to find multiple solutions in corresponding Reproducing Kernel Hilbert Spaces (RKHSs) simultaneously. To achieve this goal, multiple individual kernel losses are integrated into a unified ensemble loss. Therefore, each model can co-optimize to learn its optimal parameters by minimizing a unified ensemble loss in multiple RKHSs. Furthermore, we apply our proposed ensemble loss into the deep network paradigm and take the sub-network as a kernel mapping from the original input space into a feature space, named Deep-MKEL (D-MKEL). Our D-MKEL model can utilize the diversified deep individual sub-networks into a whole unified network to improve the classification performance. With t...

Weighted Neighborhood Preserving Ensemble Embedding

Electronics, 2019

Neighborhood preserving embedding (NPE) is a classical and very promising supervised dimensional ... more Neighborhood preserving embedding (NPE) is a classical and very promising supervised dimensional reduction (DR) technique based on a linear graph, which preserves the local neighborhood relations of the data points. However, NPE uses the K nearest neighbor (KNN) criteria for constructing an adjacent graph which makes it more sensitive to neighborhood size. In this article, we propose a novel DR method called weighted neighborhood preserving ensemble embedding (WNPEE). Unlike NPE, the proposed WNPEE constructs an ensemble of adjacent graphs with the number of nearest neighbors varying. With this graph ensemble building, WNPEE can obtain the low-dimensional projections with optimal embedded graph pursuing in a joint optimization manner. WNPEE can be applied in many machine learning fields, such as object recognition, data classification, signal processing, text categorization, and various deep learning tasks. Extensive experiments on Olivetti Research Laboratory (ORL), Georgia Tech, C...

Download

Structure injected weight normalization for training deep networks

Multimedia Systems

Weight normalization (WN) can help to stabilize the distribution of activations over layers, whic... more Weight normalization (WN) can help to stabilize the distribution of activations over layers, which boost the performance of DNNs in generalization. In this paper, we further propose deep structural weight normalization (DSWN) methods to inject the network structure measurements into the WN to fully acknowledge the data propagation through the neural network. In DSWN, two novel structural measurements are developed to impose regularity on each network weight using different penalty matrices. One is sparsity measurement (DSWN-SM). In this measurement, L1,2 weight regularization is applied in our proposed model to promote competition for features between network weights to obtain a sparsity network and finally prune the network. The other is neuron measurement (DSWN-NM). It uses L2 norm of column weight to scale up or down the importance of each intermediate neuron, which leads to accelerating the speed of network convergence. Extensive experiments on several benchmark image datasets using fully connected network and convolution neural network are performed, and the proposed DSWN-SM and DSWN-NM methods are compared with state-of-the-art sparsity and weight normalization methods. The results show that DSWN-SM can reduce the number of trainable parameters while guaranteeing high accuracy, whereas DSWN-NM can accelerate the convergence while improving the performance of deep networks.

Incomplete-Data Oriented Dimension Reduction via Instance Factoring PCA Framework

In this paper, we propose an instance factoring PCA (IFPCA) framework for dimension reduction in ... more In this paper, we propose an instance factoring PCA (IFPCA) framework for dimension reduction in incomplete datasets. The advantage of IFPCA over the traditional PCA is that, a penalty is imposed on the instance space via a scaling-factor to suppress the effect of outliers in pursuing projections. We geometrically use two scaling-factor strategies, total distance and cosine similarity metrics. Both strategies can learn the relationship between each data point and the principal projection in the feature space. In this way, better low-rank projections are obtained through scaling the data iteratively to suppress the impact of noise in the training set. Extensive experiments on COIL-20, ORL and USPS datasets prove the superiority of the proposed framework over state-of-the-art dimensionality reduction methods such as LSDA, gLPCA, RPCA-OM, PCA, LPP and RCDA.

Robust low rank representation via feature and sample scaling

Neurocomputing, 2020

Low-rank representation (LRR) is a very competitive technique in various real-world applications ... more Low-rank representation (LRR) is a very competitive technique in various real-world applications for its powerful capability in discovering latent structure of noisy or corrupted data set. However, traditional low-rank models treat each data point and feature equally so that noisy data cannot be detected and suppressed effectively and have obvious deterioration in performance, especially in heavy noisy scenario. In this paper, to address this problem, we develop a method of feature and sample scaling for low rank representation. The importance of data points and their features in both feature and sample spaces are considered, as such, clean data points and noisy data points and their features can be distinguished. In addition, based on the observation that noisy data points are usually deviated far away from the principal projection of the data set, a cosine similarity metric between data vector and the principal projection vector is developed to measure the importance of each sampl...

TrustWalker: An Efficient Trust Assessment in Vehicular Internet of Things (VIoT) with Security Consideration

Sensors

The Internet of Things (IoT) is a world of connected networks and modern technology devices, amon... more The Internet of Things (IoT) is a world of connected networks and modern technology devices, among them vehicular networks considered more challenging due to high speed and network dynamics. Future trends in IoT allow these inter networks to share information. Also, the previous security solutions to vehicular IoT (VIoT) much emphasize on privacy protection and security related issues using public keys infrastructure. However, the primary concern about efficient trust assessment, authorized users malfunctioning, and secure information dissemination in vehicular wireless networks have not been explored. To cope with these challenges, we propose a trust enhanced on-demand routing (TER) scheme, which adopts TrustWalker (TW) algorithm for efficient trust assessment and route search technique in VIoT. TER comprised of novel three-valued subjective logic (3VSL), TW algorithm, and ad hoc on-demand distance vector (AODV) routing protocol. The simulated results validate the accuracy of the p...

Download

An Improved Hybrid Method Combining Gravitational Search Algorithm with Dynamic Multi Swarm Particle Swarm Optimization

IEEE Access, 2019

GSA is badly suffering from slow convergence rate and poor local search ability, when solving com... more GSA is badly suffering from slow convergence rate and poor local search ability, when solving complex optimization problems. To solve this problem, a new hybrid population-based algorithm is proposed with the combination of dynamic multi swarm particle swarm optimization and gravitational search algorithm (GSADMSPSO). The proposed algorithm has divided the main population of masses into smaller sub-swarms and also stabilizing them by presenting new neighborhood strategy. Then, by adopting the global search ability of proposed algorithm, each agent (particle) improves the position and velocity. The main idea is to integrate the ability of GSA with the DMSPSO to enhance the performance of exploration and exploitation of a proposed algorithm. In order to evaluate the competences of the proposed algorithm, benchmark functions are employed. The experimental results have been confirmed a better performance of GSADMSPSO as compared to the other gravitational and PSO variants in terms of fitness rate.

Download

A New Nearest Centroid Neighbor Classifier Based on K Local Means Using Harmonic Mean Distance

Information, Sep 14, 2018

The K-nearest neighbour classifier is very effective and simple non-parametric technique in patte... more The K-nearest neighbour classifier is very effective and simple non-parametric technique in pattern classification; however, it only considers the distance closeness, but not the geometricalplacement of the k neighbors. Also, its classification performance is highly influenced by the neighborhood size k and existing outliers. In this paper, we propose a new local mean based k-harmonic nearest centroid neighbor (LMKHNCN) classifier in orderto consider both distance-based proximity, as well as spatial distribution of k neighbors. In our method, firstly the k nearest centroid neighbors in each class are found which are used to find k different local mean vectors, and then employed to compute their harmonic mean distance to the query sample. Lastly, the query sample is assigned to the class with minimum harmonic mean distance. The experimental results based on twenty-six real-world datasets shows that the proposed LMKHNCN classifier achieves lower error rates, particularly in small sample-size situations, and that it is less sensitive to parameter k when compared to therelated four KNN-based classifiers.

Download

Hybrid self-inertia weight adaptive particle swarm optimisation with local search using C4.5 decision tree classifier for feature selection problems

Connection Science

ABSTRACT Feature selection is an important task to improve the classifier’s accuracy and to decre... more ABSTRACT Feature selection is an important task to improve the classifier’s accuracy and to decrease the problem size. A number of methodologies have been presented for feature selection problems using metaheuristic algorithms. In this paper, an improved self-adaptive inertia weight particle swarm optimisation with local search and combined with C4.5 classifiers for feature selection algorithm is proposed. In this proposed algorithm, the gradient base local search with its capacity of helping to explore the feature space and an improved self-adaptive inertia weight particle swarm optimisation with its ability to converge a best global solution in the search space. Experimental results have verified that the SIW-APSO-LS performed well compared with other state of art feature selection techniques on a suit of 16 standard data sets.

Wormhole attack on On Demand routing protocols in mobile ad hoc networks

A mobile ad hoc network (MANET) consists of mobile wireless nodes. In MANET nodes are selfmotivat... more A mobile ad hoc network (MANET) consists of mobile wireless nodes. In MANET nodes are selfmotivated topologies can randomly change their geographic locations. Wireless networks are susceptible to many attacks, including an attack known as the wormhole attack. On occurrence of wormhole can cause a significant breakdown in communication across a wireless network. This paper explores and compares the effect of wormhole attacks on the performance of on demand routing protocols in mobile ad hoc network (MANET). The evaluation has been done by studying and comparing end to end delay, packet delivery ratio and throughput for wormhole attack and without wormhole attack. Keywords— Mobile ad hoc networks; routing protocol; wormhole attacks; end to end delay; throughput; packet delivery ratio.

Download

Sumet Mehta

Uploads

Papers by Sumet Mehta

Log In