Online ISSN: 2515-8260

Keywords : machine learning


An Ensemble Framework Based Outlier Detection System in High Dimensional Data

N Jayanthi; Dr Burra Vijaya Babu; Dr N Sambasiva Rao

European Journal of Molecular & Clinical Medicine, 2020, Volume 7, Issue 4, Pages 1162-1175

Machine learning based outlier detection methods are widely used in various domains. However, an ensemble of such detection methods could leverage detection performance. The existing ensemble methods made up of multiple unsupervised learning algorithms lack in ideal strategy for choosing right candidates as constituent detectors. It resulted in mediocrity in model stability and accuracy. To overcome this problem, in this paper, we propose an ensemble framework based outlier detection system in high dimensional data. It has ideal mechanism for effectively choosing base outlier detectors. Out of many candidate outlier detectors, the ones that yield highest performance are combined. An algorithm named Average Selection and Ensemble of Candidates for Outlier Detection (ASEC-OD). Many real world datasets are used for empirical study. The results of experiments revealed that the proposed framework outperforms many existing methods

Identification and Detection of Abnormal Human Activities using Deep Learning Techniques

ASHISH SHARMA; NEERAJ VARSHNEY

European Journal of Molecular & Clinical Medicine, 2020, Volume 7, Issue 4, Pages 408-417

In recent years, it is in public to use the surveillance cameras for continuous monitoring of public and private spaces because of increasing crime. Most current surveillance systems need a human operator to constantly watch them and are ineffective as the amount of video data is increasing day by day. Surveillance cameras will be more useful tools if instead of passively recording; they generate warnings or real-time actions when unusual activity is detected. But recognizing and classifying human activity as normal or abnormal from a live video stream is a stimulating job in the pitch of CPU vision. There is a need for a smart surveillance system for the automatic identification of abnormal behaviour of humans for a specific-scene. Presentpaperstretches an overview of different machine learning methods used in recent years to develop such a model. It also gives an exposure to the recent works in the field of anomaly detection in surveillance video and its applications

Prediction of the Crop Cultivating using Resembling and IoT Techniques in Agricultural Fields for Increasing Productivity

Anant Ram; Rakesh Kumar

European Journal of Molecular & Clinical Medicine, 2020, Volume 7, Issue 4, Pages 50-53

The agriculture plays a prevailing job in the development of the nation's economy. Atmosphere and other natural changes has become a significant danger in the agribusiness field. AI is a fundamental methodology for accomplishing viable and viable answers for this issue. Harvest Prediction includes anticipating the best output from accessible authentic information like climate parameters and soil parameters. This recommender system uses real time data as input to the machine learning. The sensors collect data from the soil and send that data to the cloud (firebase). Then the machine learning model retrieves that data and predicts the best crop and sends that crop to the cloud. We develop an android application which retrieves the sensor values from the cloud and displays them. This forecasting facilitates the farmer to forecast the best crop earlier than cultivating onto the agriculture field, which in turn increases the productivity.

Data Coloring in Trusted and Ambiguous Cloud Computing using Sheltered Possessions Machine Learning Technique

Ellappan Venugopal; Dr Rajesh Thumma; Dr. R. Sreeparimala; Dr.Anusha K; Dr. T. Thulasimani

European Journal of Molecular & Clinical Medicine, 2020, Volume 7, Issue 2, Pages 3270-3276

Confidence and security prevent businesses from fully embracing cloud platforms. To protect the
clouds, providers must first secure virtualized data-center resources, uphold user privacy, and protect data
integrity. The authors suggest that the trust overlay network be used through multiple data centers to implement
a reputation system to build trust between service providers and data owners. Data coloring and software
watermarking methods protect shared data objects and highly distributed software modules. These strategies
protect multi-path authentication, enable a single sign-on in the cloud, and tighten access control for sensitive
data in the public and private clouds. Protection against tampering is tamper proofing, so unauthorized changes
to the software (for example, removing a watermark) can lead to passive code. We will briefly examine the
technology available for each type of protection. P2P technology opens our work to low cost copyrighted
content delivery. The advantages are mainly delivery cost, high content availability and copyright compliance in
exploring P2P network resources.

Stock Prediction using Sentiment analysis and Long Short Term Memory

Harsh Panday; V. Vijayarajan; Anand Mahendran; A. Krishnamoorthy; V.B. Surya Prasath

European Journal of Molecular & Clinical Medicine, 2020, Volume 7, Issue 2, Pages 5060-5069

The Stock market is a shambolic place for prediction as there are plenty of factors that affect the stock market simultaneously. Numerous studies have been conducted regarding this field, in hopes that one day accurate stock values can be predicted. This paper introduces a hybrid algorithm that incorporates Twitter sentiment analysis and Long Short Term Memory to predict next day closing values of a stock. Our proposed algorithm exploits the temporal correlation between public sentiment and its effect on stock values. We use Part-of-speech tagging to perform sentiment analysis and Long Short Term Memory for foretelling the next day closing price of the stock, both of these combined gives us a decent picture regarding the future of the stock.

Investigation and development of machine Learning Challenges in Video Interviews

DILIP KUMAR SHARMA; ASHISH SHARMA

European Journal of Molecular & Clinical Medicine, 2020, Volume 7, Issue 4, Pages 425-433

This paper audits and talks about examination propels on "logical AI" in PC vision. We centre on a specific zone of the "Seeing People" (LAP) topical space: primary imitations and character investigation. Our point is to variety the computational knowledge and PC vision networks mindful of the significance of creating logical systems for PC helped dynamic applications, for example, robotizing enlistment. Decisions dependent on character attributes are being made routinely by human asset offices to assess the up-and-comers' ability of social inclusion and their capability of profession development. Be that as it may, deducing character attributes and, as a rule, the procedure by which we people structure a first impression of individuals, is profoundly emotional and might be one-sided. Past investigations have shown that knowledge machineries can figure out how to imitate human choices. In this paper, we go above and beyond and figure the issue of clarifying the choices of the models as methods for distinguishing what visual perspectives are significant, seeing how they identify with choices recommended, and potentially picking up knowledge into unfortunate negative inclinations. We structure another test on reasonableness of knowledge machineries for first impressions examination. We portray the setting, situation, assessment measurements and starter results of the opposition. Supposedly this is the first exertion regarding difficulties for logic in PC vision. Moreover, our test configuration involves a few other measurable and subjective components of oddity, including a "competition" setting, which joins rivalry and coordinated effort.

An Analysis of Machine Learning Algorithms in Profound

G. Siddarth .; Dr. Preethi Nanjundan

European Journal of Molecular & Clinical Medicine, 2020, Volume 7, Issue 2, Pages 5179-5192

AI is the usage of man-made consciousness (AI) that enables systems to subsequently to take in and improve as a matter of fact without being unequivocally modified. AI focuses on the improvement of PC programs that can get to data for learning itself. Machine Learning tasks are classified into supervised, unsupervised, and reinforcement learning. In this paper, we discuss in-depth comparisons of all supervised and unsupervised algorithm.

A REVIEW OF MACHINE LEARNING FRAMEWORKS FOR EARLY AND ACCURATE PREDICTION OF NEOADJUVANT CHEMOTHERAPY RESPONSES

Uddaraju Susmitha; Narasingarao, M. R

European Journal of Molecular & Clinical Medicine, 2020, Volume 7, Issue 4, Pages 1040-1050

The ability to predict the reaction of breast tumors to neoadjuvant chemotherapy from the get-go over the span of treatment can delineate patient’s dependent on the reaction for explicit tolerant treatment procedures. From now on, reaction to neoadjuvant chemotherapy is measured as being based on physical examination or breast imaging (mammogram, mri, or normal MRI). There is a powerless connection with these projections and with the actual tumor size as measured by the pathologist through authoritative procedure. Given the numerous options open to Neoadjuvant chemotherapy (NAC), it is important to develop a plan to predict response over the care period. Sadly, as long as certain people are not seen as responding, their condition can never again be specifically resectable, so this situation should be preserved at a strategic remove from progressing response appraisal protocols throughout the care regimen. This paper provides a review of all the existing frameworks of machine learning involved to perform accurately neoadjuvant chemotherapy responses

Identification and Prediction of Liver Disease using Logistic Regression

Neeraj Varshney; Ashish Sharma

European Journal of Molecular & Clinical Medicine, 2020, Volume 7, Issue 4, Pages 106-110

Identification of disease at a beginning stage is very essential for higher treatment. It’s a awfully complicated task for medical researchers to predict the illness within the early stages because of delicate symptoms. Typically the symptoms turn out to be evident once it's too late. to beat this issue, this project aims to boost disease designation victimization machine learning approaches. The most objective of this analysis is to use categorization techniques to spot the liver patients from healthy people. This project conjointly aims to match the categorization techniques supported their presentation factors. To serve the medical community for the designation of disease between patients, a graphical computer interface is urbanized victimization python (Node RED). The GUI will be promptly used by doctors and medical practitioners as a screening tool for the disease.

An Efficient Reactive Join Nested Loop Machine Learning Inputs In Autonomous Smart Grid Environment

Mr. Nilesh; Dr. M. Prasad; Dr.R. Sabitha; Dr Raghavender K V; Dr Varun Gupta

European Journal of Molecular & Clinical Medicine, 2020, Volume 7, Issue 7, Pages 211-218

Adaptive join algorithms have recently attracted a lot of attention in emerging applications that provide data through autonomous data sources in diverse network environments. Their main advantage over traditional joining technologies is that they can give the results they join as soon as the first input duplex is available, thereby improving the pipeline by joining the results and hiding the source or network delay. In this paper, we first suggest the new adaptive two-way join algorithm DINER (Dual Indexed-Loops Reactive Join) to increase the result rate. Diner combines two main components: the novel Retrofit technology, which allows algorithms to quickly switch between memory processing and a clea
r flushing approach aimed at increasing the productivity of memory tuples in producing results in the stage of reaching online. We are expanding the application of specific technology for a more challenging setup: managing more than two inputs. The Multi Active Relational Join Algorithm (MARA) is a multi-path joint operator that claims its principles from DINER. Tara surpasses previous compatible joint algorithms in our experiments with real and synthetic data sets, makes the best use of available memory and produces duplicates of results at significantly higher rates. In the presence of multiple experiments, our experiments show that MARA can produce a high percentage of initial results and surpass existing technologies for adaptive multi-path joining.

SKIN DISEASE DETECTION USING COMPUTER VISION AND MACHINE LEARNING TECHNIQUE

Leelavathy S; Jaichandran R; Shobana R; Vasudevan .; Sreejith S Prasad; Nihad .

European Journal of Molecular & Clinical Medicine, 2020, Volume 7, Issue 4, Pages 2999-3003

Skin types of diseases are most common among the globe, as people get skin disease due to inheritance, environmental factors. In many cases people ignore the impact of skin disease at the early stage. In the existing system, the skin disease are identified using biopsy process which is analyzed and medicinal prescribed manually by the physicians. To overcome this manual inspection and provide promising results in short period of time, we propose a hybrid approach combining computer vision and machine learning techniques. For this the input images would be microscopic images i.e histopathological from which features like color, shape and texture are extracted and given to convolutional neural network (CNN) for classification and disease identification. Our objective of the project is to detect the type of skin disease easily with accuracy and recommend the best and global medical suggestions.
This paper proposes a skin disease detection method based on image processing and machine learning techniques. The patient provides an image of the infected area of the skin as an input to the prototype. Image processing techniques are performed on this image and feature values are extracted and the classifier model predicts the disease. The proposed system is highly beneficial in rural areas where access to dermatologists are limited. For this proposed system, we use Pycharm based python script for experimental results.

An Empirical Study of Deep Learning Strategies for Spatial Data Mining

K. Sivakumar; A.S. Prakaash

European Journal of Molecular & Clinical Medicine, 2020, Volume 7, Issue 2, Pages 5124-5132

The emergence of scalable frameworks for machine learning to efficiently analyse and derive valuable insights from these data has triggered growing volumes of data collected. Huge spatial data frameworks cover a wide variety of priorities, including tracking of infectious diseases, simulation of climate change, etc. Conventional mining techniques, especially statistical frameworks to handling these data, are becoming exhausted due to the rise in the number, volume and quality of spatio-temporal data sets. Various machine learning tasks have recently shown efficiency with the development of deep learning methods. We therefore include a detailed survey in this paper on important impacts in the application of deep learning techniques to the mining of spatial data.

Sentimenatl Analysis To Improve Teaching And Learning

K. Sai Tulasi; N. Deepa

European Journal of Molecular & Clinical Medicine, 2020, Volume 7, Issue 3, Pages 2194-2200

Today internet is playing a vital role in the modern world. From young aged children to old aged people are utilizing the technology. By the use of this technology to determine the mental taught of the student towards the learning environment the sentimental analysis method is performed. For this they created a separate portal for the student which has unique ID and password for individual one. Inside the web portal it consists of various sets of question and it records the performance of the student towards it. The feedback can be subdivided into three categories mainly positive, negative, and neutral. By using the feedback they can predict the sentiments of the student. By using the machine learning algorithm they can predict the sentiment of the students in the learning aspects. Student sentiment has been get distinguished by the use of the polarity. This is can uses the random forest algorithm. The random forest algorithm can provides the accuracy of about 92% with effective outcome.

A Pilot Survey Of Machine Learning Techniques In Smart Grid Operations Of Power Systems

Dr. Shafali Jain; Ravi Prasad. B; Dr. C. Ashok kumar; Dr Mohan Dattu Sangale; E. Fantin Irudaya Raj

European Journal of Molecular & Clinical Medicine, 2020, Volume 7, Issue 7, Pages 203-210

The smart grid discusses to next generation power grids, with multi-directional streams of electricity and evidence to make aextensive distributed network. Through smart grid, the power system converts smart by communicating, sensing, control and applying intelligence. For superlative system, the smart grid technologies are furthercompanionable to certify many roles which can elevate with the amalgamation of the use of substance generation and transmission. The Smart grid is also kept the environment free from pollution; diminish the cost, effective operations, against all categories of threats and danger. Machine learning process isthe calculations which help in information handling to discover concealed examples or the forecast of results. The target of this archive is to look at the most utilized strategies of machine learning, for example, Vector Machine Backing, Descriptive Discriminant Analysis, Decision Trees what's more, Neural Networks, in Smart Grid applications. To this end, an examination is done in important distributions of the current writing.

MAGNETIC RESONANCE MACHINE LEARNING METHOD FOR PREDICTING GEO GRAPHICAL LOCATION SPECIFICATION

Sudhir Sharma; G Shobana; 3L Chandra Sekhar Reddy; P Madhuri; P Naveen

European Journal of Molecular & Clinical Medicine, 2020, Volume 7, Issue 2, Pages 3226-3233

Graph theory is a branch of discrete mathematics that deals with the connections among
entities. It has been proven to be a very beneficial and powerful mathematical tool and has a
wide range of applications to handle complex problems in various domains. The aim of this
work is two folds: first, to understand the basic notion of graph theory and second, to
emphasis the significance of graph theory through a real-time application used as a
representational form and characterization of brain connectivity network, as is machine
learning for classifying groups depending on the features extracted from images. This
application uses different techniques including preprocessing, correlations, features or
algorithms. This paper illustrates an automatic tool to perform a standard process using
images of the Magnetic Resonance Imaging (MRI) machine. The process includes preprocessing,
building the graph per subject with different correlations, atlas, relevant feature
extraction according to the literature, and finally providing a set of machine learning
algorithms that can produce analyzable results for physicians or specialists. Further, to
demonstrate the importance of graph theory, this article addresses the most common
applications for graph theory in various fields.

SPAM DETECTION OF PHISHING WEBSITES USING ML

Dr. J. Selvakumar; Mr. R. Prithiviraj; Mr. Joshua Jafferson; Mr.S. Bashyam

European Journal of Molecular & Clinical Medicine, 2020, Volume 7, Issue 4, Pages 2184-2190

In today’s internet era various websites through which a number of individuals purchase items. There are certain online forums which request their users to provide confidential data such as card number, cvv, pin number etc. for various malicious practices. These websites are referred as Phishing Websites. Therefore, to distinguish between the authentic website and the malicious website we suggested an intelligent, adaptable, and efficient model that utilizes Machine learning techniques. We carry through the project using the algorithm of classification and different methods to gather the phishing websites dataset to verify its validity. These spoofing websites are differentiated on certain significant attribute such as encryption standards, Domain Identity, URL and security. The project will utilize machine learning concept thus informing the user if the website is legal or not. This software is highly secured and can be utilized by many E-commerce ventures so as to provide hassle free transaction. Machine Learning design utilized in the project gives good results when compared with other standard classification algorithms. Detection of Phishing web site is ML intelligent and effective model that’s supported victimization classification or association data processing algorithms. The algorithms we are using here is logistic regression. We are also using decision tree classifier so that we can make a point-to-point comparison between them which will help us to know parameters like accuracy and time taken.

Analysis on Worldwide Online Rating Systems by Using Multi Linear Regression Technique in Machine Learning

Shashishekhar .; Ashish Sharma

European Journal of Molecular & Clinical Medicine, 2020, Volume 7, Issue 4, Pages 111-115

In today’s world online rating systems are extensively used for making their decisions on a particular product or item on the web. For getting benefit, some people are always trying to manipulate corresponding systems by giving unreasonable ratings. So, resolving true ratings of such products comes extremely important and it is a crucial problem. Fake reputation is the problem that is being occurred by unreasonable ratings. In this paper we propose Multi linear regression algorithm in machine learning which will give correct reputation by eliminating unreasonable ratings

An approach for smart cities parking based on cloud computing and Machine Learning by Genetic Ant Colony Algorithm

SAURABH SINGHAL; NARENDRA MOHAN

European Journal of Molecular & Clinical Medicine, 2020, Volume 7, Issue 4, Pages 434-441

This research deals with the major problem that we face in major part of India and across most of the world. The research deals with the main problem of traffic congestion and road accidents that is basically caused because of the improper parking management. So, it is mandatory for all the cities to have a well managed parking system. However, in the past many researches has been conducted to propose an solution that leads to suitable smart paring algorithm. On reading more about the researchers conducted in the past, it was clear that each research has its own pros and cons. This paper reflects on the research conducted to design an algorithm that leads to a cloud based smart algorithm that is secure and is convenient enough to develop a system that can be used to manage the available slots and can notify the users about the available parking slot beforehand to the client. The paper also focuses on the result analysis part that clearly shows that the algorithm designed is more accurate than other algorithms used in the past. We have designed our algorithm using ACO, decision tree, and GPS mapping over cloud. The idea of working on this research was to provide a solution that is cost effective, helps people on large scale and maintains the laws and order.

Students Attention and Engagement Prediction Using Machine Learning Techniques

Leelavathy S; Jaichandran R; Shantha Shalini K; Surendar B; Aswin K Philip; Dekka Raja Ravindra

European Journal of Molecular & Clinical Medicine, 2020, Volume 7, Issue 4, Pages 3011-3017

Real-time student engagement tracking is an important step towards education. Current approach doesn’t consider student engagement detection using biometric features. In this project, we propose a hybrid architecture invoking student’s eye gaze movements, head movements and facial emotion to dynamically predict student attention and engagement level towards the tutor and based on the output value the content is changed dynamically. Hence this concept has a huge scope in e-learning, class room training, analyse human behaviour. This project covers main process like Eye Ball, facial emotion and head movements Human Beings. For feature extraction step, we used Principal Component Analysis (PCA) for facial emotion recognition, Haar Cascade for pupil detection and Local Binary Patterns for recognizing head movements and OpenCV for machine learning model generation and comparison.

TECHNOLOGY PERCEPTIVE OF INTELLIGENT MACHINE LEARNING DATA ANALYTICS METHOD USING FUZZY LOGIC SYSTEMS

Dr.Syed Khasim; Dr.T. Thulasimani; Bottu Gurunadha Rao; Dr.S.V. Sudha; M P Rajakumar

European Journal of Molecular & Clinical Medicine, 2020, Volume 7, Issue 2, Pages 3241-3249


The instinct of human beings is very complex. The understanding of this instinct needs a strong
dimensional analysis of the knowledge of discourse. The computer systems are now trained to realize
how things operate in a real-time environment for intelligent analysis. This effort although very
progressive has a restriction. There is an intelligence gap which makes human one step above the
machine. The fuzzy logic system can be employed to make a machine understand this intelligence gap
in a better way. In other words, fuzzy logic is a computational Intelligence technique that makes a
computer understand and think the way humans do. The fuzzy logic system is now attracting
scientists and engineers around the world since by integrating its abilities with soft computing
techniques like neuro and chaos computing, genetic algorithm, probability reasoning, and immune
networks, it can handle the problems that had not been solved before. Not only has the fuzzy system
significantly enhanced knowledge-based or expert system technology, but it has also primarily
changed the granularity of intelligence. For example, with the help of the fuzzy logic system,
industrialists of home appliances are today embedding intelligence in specific productsThe aim of this
study is two folds: first, to understand the fuzzy logic system for effective decision making and
second, to demonstrate the presence of this intelligence gap through real-time examples. The
examples are selected cautiously to illustrate and demonstrate the applications of the fuzzy logic
system for every reader.

Moving Towards Non-AI To AI

Nargis A Vakil; S.B. Goyal

European Journal of Molecular & Clinical Medicine, 2020, Volume 7, Issue 2, Pages 5638-5646

A large number of researches have been conducted in the field of AI. This paper is all about the enhancements made in this popular field. Making a machine that is able to understand the background ideas of the words is very essential as it can increase the chances of better translation as well as can execute conversations as humans do. In particular, this paper states the difference between the AI and the Non-AI tasks. The work is generated for new candidates coming in the area of AI as well as some issues related to AI are also talked about.

Prediction of Population Growth using Machine Learning Techniques

Brintha Rajakumari S; Padmanabhan P; Christy S; Nandhini M

European Journal of Molecular & Clinical Medicine, 2020, Volume 7, Issue 4, Pages 1885-1890

Population growth prediction shows the future rate of fertility, mortality and migration of people of a country. It is very important for the population and health system. Nowadays, Machine learning concepts are most growing and popular for predicting future values. In order to predict population growth, the machine learning concept applied to build the map between year and population growth. The paper investigates the population growth of Indian government population data using time series forecasting machine learning techniques and analyzed byLinear regression, Support Vector Regression, Multilayer perceptron and Decision tree classifier. The optimum prediction method is based on the technique which gives very less error rate. The increment or degradation of instances in datasets do not affect the performance of the techniques is also analysed. The obtained result shows that the linear regression gives less error than the other classifier to predict population growth of India.

The Advent of Artificial Intelligence in Cardiology: The Current Applications and Future Prospects

K Prashanth; M Manjappa; C Srikar

European Journal of Molecular & Clinical Medicine, 2020, Volume 7, Issue 1, Pages 14-20

The technology of artificial intelligence is emerging as a promising entity in cardiovascular medicine, with the potential to improve diagnosis and patient care. In this article we review the literature on artificial intelligence and its utility in cardiology. We provide a detailed description of concepts of artificial intelligence tools like machine learning, deep learning and cognitive computing. This review discusses the current evidence, applications, future prospects and limitations of artificial intelligence in cardiology.