International Journal of Computer (IJC - Global Society of Scientific Research and Researchers, GSSRR)
Not a member yet
459 research outputs found
Sort by
Bank Credit Risk Management Using Machine Learning Algorithms
Get PDFPrior PCs was simply sorted as a need of an individual yet now it turns into a need of a person. AI fills in as a significant part in field of PC. Machine can\u27t thoroughly consider various circumstances however it can draw diverse kind of connections between various highlights and qualities. The significant piece of our life is to stay away from false exercises yet till now we can\u27t authority over it. Credit business is one of the significant organizations of business banks. Deceitful exercises can be handle through installing AI calculations in our everyday life. In this venture we will utilize directed AI and for that we need to give named information to the AI calculation. This paper centers around anticipating SME client status for time of a half year by using application scoring extra to client conduct highlights. By using Neural Networks, Support Vector Machines and Inclination Boosting, execution examination and furthermore highlight investigation for client conduct are directed
Overlapping Community Detection using Local Seed Expansion
Get PDFCommunities are usually groups of vertices which have higher probability of being connected to each other than to members of other groups. Community detection in complex networks is one of the most popular topics in social network analysis. While in real networks, a person can be overlapped in multiple communities such as family, friends and colleagues, so overlapping community detection attracts more and more attention. Detecting communities from the local structural information of a small number of seed nodes is the successful methods for overlapping community detection. In this work, we propose an overlapping community detection algorithm using local seed expansion approach. Our local seed expansion algorithm selects the nodes with the highest degree as seed nodes and then locally expand these seeds with their entire vertex neighborhood into overlapping communities using Personalized PageRank algorithm. We use F1_score( node level detection ) and NMI( community level detection ) measures to assess the performances of the proposed algorithm by comparing the proposed algorithm’s detected communities with ground_truth communities on many real_world networks. Experimental results show that our algorithm outperforms over other overlapping community detection methods in terms of accuracy and quality of overlapped communities
Human Resource Information System with Digital Archiving
Get PDFThis study aimed to develop an automated tool for Human Resource Information System (HRIS) with security code and verifier integrated module. Rapid Application Development (RAD) Model was used in the planning, creating, deploying, and testing the system. VisualBasic.net, Navicat, and Dezign were utilized in the system development and MySQL as database. The system helps manages employees’ records, in particular, information for leave credits, service records, and training development programs. It also tracks employees\u27 performance and skills and manage the office resources. Using the system evaluation based on the ISO 9126 standard, the system has a high rate of usability (4.27), functionality (4.35), maintainability (4.23), and efficiency (4.30). Thus, the system is believed to provide a significant contribution to the productivity of the Human Resource employees; thereby, generating a due and timely feedback to the administration. 
CNN Transfer Learning for Automatic Fruit Recognition for Future Class of Fruit
Get PDFDeep fruit recognition model learned on big dataset outperform fruit recognition task on difficult unconstrained fruit dataset. But in practice, we are often lack of resources to learn such a complex model, or we only have very limited training samples for a specific fruit recognition task. In this study we address the problem of adding new classes to an existing deep convolutional neural network framework. We extended our prior work for automatic fruit recognition by applying transfer learning techniques to adding new classes to existing model which was trained for 15 different kind of fruits. Pre-trained model was previously trained on a large-scale dataset of 44406 images. To add new class of fruit in our pre-trained model, we need to train a new classifier which will be trained for scratch, on the top of pre-trained model so, that we can re- purpose the feature learned previously for the dataset. Transfer learning using our pre-trained model has been demonstrated to give the best classification accuracy of 95.00%. The experimental results demonstrate that our proposed CNN framework is superior to the previous state-of-the- art networks
Enhanced Wi-Fi Security of University Premises Using MAC Address and Randomly Generated Password
Get PDFMany solutions are available for setting up wireless home networks to get internet connectivity working as quickly as possible. It is also quite risky as numerous security problems can result. Today’s Wi-Fi networking products do not always help the situation as configuring their security features, and they can be time-consuming. In this paper, an improved security protocol is proposed for University premises, which is a combination of the process of MAC address filtering and random password generation. If the MAC address match, then the server will send a randomly generated password to the client. As a result, the whole network will face fewer intruders, and the security will be of top-notch. The proposed security solution was compared with the existing four security methods. The proposed solution has universality as the device and software needed for it is available all over the world
Crowdsourced Machine Learning Based Recommender for Software Design Patterns
Get PDFSoftware technology has become an essential part of human lives today. The role of software Engineers in making this technology as success is very fundamental. In software Engineering, the toughest stage is to design software as there is no particular rule or formula to covert requirements into design representation. A designer designs software using skills, critical thinking ability and previous experience only. To make this process easy, the design patterns came into existence which are the solutions that can be used repetitively to solve design problems. There have been several pieces of research presented regarding design Patterns but it is hard to find research regarding how the patterns are perceived and used in industries today and what nature of application uses which specific patterns. This paper uses a crowdsourced approach to acquire the finest practices that are being used in industries today including which quality attributes are affected most by the implementation of these patterns and which patterns are suitable for what type of applications. It also uses a machine learning supervised algorithm (Matchbox Recommender) to predict suitable design pattern for different nature of applications
Prediction of Soil Macronutrients Using Machine Learning Algorithm
Get PDFIn this research work, machine learning algorithms were applied to find the relationship between independent variables and dependent variables for soil data analysis. The independent variables include moisture, temperature, soil pH, Cation Exchange Capacity(CEC) whereas, the dependent variables include Nitrogen, Phosphorus and Potassium (NPK). This research concludes relationships between Phosphorus, Potassium, soil pH and CEC; Nitrogen and soil moisture and temperature using machine learning(ML) algorithms so as to deduce NPK content of soil. A comparative analysis with obtained results from each ML method is also presented. Machine learning algorithms are best performed on data with multiple independent variables. The values computed for nitrogen relationship were more accurate than PK relationship values. The accuracy of data set I was less than data set II. A large data set would produce more accurate results for both data sets
Infection Severity Detection of CoVID19 from X-Rays and CT Scans Using Artificial Intelligence
Get PDFDecember 2019, marked with a widespread infection due to a new matured member of SARs Virus named as SARS-CoV2 (Novel Corona Virus-2019) infecting more than 20 lakhs people across the globe. This effect made the World Health Organization to declare COVID-19 (Corona Virus Disease, 2019) as a pandemic situation and called a worldwide lockdown to dampen and flatten the infectious curve and diminish the infection growth. With Limited number of COVID-19 test kits in hospitals and the increasing daily cases has asked for an immediate measure for the development towards the Automatic COVID-19 Detection and Alternative Diagnosis Systems (ACD-ADS). This research presents a two-staged DenseNet architecture to diagnose the COVID19 infections from X-rays and CT-scans images to decrease the turnaround time of the doctors and check more patients during that point of time. This research work talks about the end to end solution for the diagnosis to extract and mark the most infectious regions on the imaging pictures to help the doctors and medical practitioners in this pandemic situation. The system achieved an accuracy of 99% and specificity of 94.1% using the DenseNet network on the X-rays images and an accuracy of 87% and specificity of 86.5% for the CT Scans in the Validation Sets. In a sample of 22 images for the CT-Scans of the reported patients having the COVID-19 infections in a real-time analysis, the model performed with detecting correctly for all the 22 patients. Any model can never replace a doctor nor can decide like a doctor who takes many other factors into the account that impacts a decision at a particular point of time. Hence, I propose a network called Automatic Diagnostic Medical Analysis for the COVID-19 Detection System (ADMCDS) that takes the images and tries to find the infectious regions to help the doctor better identifying the diseased part if any
COVID-19 Outbreak Data Analysis and Prediction Modeling Using Data Mining Technique
Get PDFNowadays, sustainable development is considered a key concept and solution in creating a promising and prosperous future for human societies. Nevertheless, there are some predicted and unpredicted problems that epidemic diseases are real and complex problems. Hence, in this research work, a serious challenge in the sustainable development process was investigated using the classification of confirmed cases of COVID-19 (new version of Coronavirus) as one of the epidemic diseases. Hence, the data mining predictive modeling method of data handling and predictive or forecasting the spread of COVID-19 virus. This research work mainly works on predicting or forecasting by using fbprophet. Prophet it is a python library package used for forecasting time series data based on an additive model where non-linear trends are fit with yearly, weekly, and daily seasonally, plus holiday’s effect. It works best with time series that have a strong seasonal effect and several seasons of historical data. The model helps to interpret patterns of public sentiment on disseminating related health information and assess the political and economic influence of the spread of the virus
Development of an Ontology-Based Personalised E-Learning Recommender System
Get PDFE-learning has become an active field of research with a lot of investment towards web-based delivery of personalised learning contents to learners. Some issues of e-learning arise from the heterogeneity and interoperability of learning content to suit learner’s style and preferences in order to improve the e-learning environment. Hence, this paper developed an ontology-based personalised recommender system that is needed to recommend suitable learning contents to learners using collaborative filtering and ontology. A pre-test is carried out for users in order to segment them in learning categories to suit their skill level. The learning contents are structured using ontology; and collaborative filtering is used to collects preferences from many users and then recommending the highest rated contents to users. The system is implemented using JAVA programming language with Structured Query Language (MySQL) as database management system. Performance evaluation of the system is carried out using survey and standard metrics such as precision, recall and F1-Measrure. The results from the two performance evaluation models showed that the system is suitable for recommending the required learning contents to learners