Journal of Computational Intelligence and Secure Systems of Artificial Intelligence

In this study, Fuzzy Otsu is introduced, a leading segmentation method that makes use of fuzzy logic to advance the traditional thresholding approach. The fuzzy approach overcomes visual confusion and handles noisy or fade-looking images much better than the unmodified Otsu’s method which finds clear divisions between classes. The suggested process increases the difference between picture regions by maximizing the variance in fuzzy classification to select the right thresholds. Experimental results indicate that this method performs well in maintaining object edges and details in comparison with conventional thresholding. This technique enables accurate images for remote sensing and medical picture analysis.

Keywords : Segmentation, BCV, OTSU, Fuzzy Logic

Author : Ramesh Raju

Title : Image Segmentation Thresholding Method Using the Fuzzy OTSU Method

Volume/Issue : 2025;01(03)

Page No : 01-04

A Biogeography-Based Optimisation approach is introduced in this work to resolve complex optimisation problems with fuzzy load uncertainty. In systems involving power and resource distribution, the data used by traditional methods of optimisation is often unclear or imprecise. The selected strategy manages unexpected changes in workload by using fuzzy logic in the BBO architecture, ensuring better and more long-lasting results. It considers the fuzzy changes in system load when determining habitat suitability and animal movement. Experiments show that simulation performs better than traditional methods in terms of fast convergence, accurate results and how well it changes with changing conditions. This style of decision-making really excels when conditions are uncertain and things are always changing.

Keywords : Voltage profile, Biogeography-based optimization (BBO), Fuzzy load uncertainty, Distributed generation, Power loss reduction

Author : M Rahul and K Gandhi

Title : Biogeography-Based Optimization Taking Fuzzy Load Uncertainty Account

Volume/Issue : 2025;01(03)

Page No : 05-08

The goal of this work is to develop a classifier that can classify tweets on Twitter. These days, Internet users use microblogging as a communication tool. They have similar views on various facets of life. We are getting closer to the Information Explosion age as microblogging sites gain prominence. With over 400 million tweets posted every day—that's an incredible 5,700 tweets per second—Twitter is the second most popular microblogging platform. Although Twitter is widely used, there isn't a precise classifier for the tweets posted on the platform. Using a learning algorithm to create a classifier that classifies the tweets after being trained on the Twitter corpus, this study aims to separate tweets and categorize them into groups such as sports and entertainment. The user can skip the tiresome process of browsing the newsfeed and just scan the tweets with this type of classifier.

Keywords : Machine learning algorithm, Microblogging, Tweets, Stemming

Author : Deepak

Title : Classifier for Twitter Tweets

Volume/Issue : 2025;01(03)

Page No : 09-11

Soft computing is a collection of interrelated approaches that enable decision-making based on trustworthy information or the experience of experts. These days, the medical field makes extensive use of a variety of soft computing approaches, including neural networks, fuzzy logic, genetic algorithms, and hybrid systems. An algorithm for analyzing lung infections is described in this research. Developing the system architecture to determine the likely disease stage that a patient may have is the primary goal. The rule-based method is used to determine the disease's severity level. To calculate an infection level, the algorithm utilizes a Rulebase output that the user has entered.

Keywords : Rulebase diagnosis, Decision support system, Medical support system

Author : Santhosh and S Raghu Ram

Title : Applying Fuzzy Rules to the Analysis of Lung Infection

Volume/Issue : 2025;01(03)

Page No : 12-15

This work uses a random point bipartite graph to present a new genetic algorithm-based graph clustering model. The model makes use of uniformly distributed random points in the data space, and gap between these points and the test points is measured and regarded as closeness. An adjacency matrix is produced using test points and random points. Correlation coefficients are calculated using the provided bipartite graph to produce a similarity matrix. To find the cluster centers, the eigenvectors of the weighted similarity matrix's singular value decomposition are taken into account and fed into an exclusive GA model. The model's performance has been compared to current standard algorithms, and it has been tested using standard datasets.

Keywords : Singular Value Decomposition, Bipartite Graph, Cluster Validity, , Clustering, Cluster Validity, Algorithm Graph, Index Genetic

Author : Sudarshan

Title : Real Coded Genetic Algorithm for Graph Clustering Based on SVD

Volume/Issue : 2025;01(03)

Page No : 16-18

The distinct designs of Rule Based Machine Translation and Statistical Machine Translation make it difficult for a pure machine translation solution to handle all translation obstacles. Once lexical and syntactic analyzers are dealt with using Rule Based, the Expectation–Maximization algorithm is run to resolve certain uncertainties with maximum likelihood. We have designed and explained an integrated hybrid machine translation system in this paper. Fusing the most effective features from different strategies is what is pursued. In the first step, RBMT converts Arabic text to English and then EM is run on the output to improve it and create the final version. To do well and become better, the EM algorithm must be able to accurately translate a frequency from one language to another. The results prove that the new approach works better than the common Rule Based approach and the EM algorithm, as proven by the BLEU system. The survey suggests that scores from the HMT system are regularly greater than those from SMT.

Keywords : Machine Translation, Arabic-English Machine, Expectation–Maximization Translation, Hybrid Machine Translation, (EM) Algorithm

Author : Rehaman Shaik

Title : Algorithm for Arabic-English Hybrid Machine Translation Optimisation

Volume/Issue : 2025;01(03)

Page No : 19-21