An approach to Sentiment Analysis using Semantics & Context
Arushi Tetarbe and Dr. Rajni Sehgal, Amity School of Engineering & Technology, Amity University, Noida, India
In today’s world the sentiment or reaction to particular product, decision or statement has proven to be highly defining in the process of governing, strategy making in businesses or monitoring cybercrimes. Sentiment is a very personal outlook on something by an individual. If we are able to judge that based on analysis of the text that the person writes on various platforms, we will be able to Product monitoring is done by almost every business out there to help in shaping their future strategies. There are various companies that offer these features as a service also. Our model is basically a deep learning model for sentiment classification that performs the task of sentiment analysis. We are using Stanford Sentiment Treebank which contains labelled text data for training the model. In order to include semantics in our model, we convert the words into their equivalent vector representation by Stanford’s unsupervised trained word embedding model called GloVe. To train the model we use LSTM (Long Short Term Memory) which is a kind of RNN (Recurrent Neural Network). This neural network is built on Keras which in turn is built on Tensorflow. Finally, our model is able to reflect the 91.33% sentiment when tested on example sentences.
Sentiment Analysis, Recurrent Neural Networks, LSTM, GloVe, Tensorflow
Natural Language to SQL
Ankita Makker and Gaurav Nayak, Department of Computer Science and Engineering, PDPM Indian Institute Of Information Technology, Design and Manufacturing, Jabalpur, India
In this research, an intelligent system is designed for the users to access the database using natural language. It accepts natural language input and then converts it into an SQL query. Using query language for dealing with databases has always been a professional and complex problem. The system currently handles single sentence natural language inputs and concentrates on MySQL database system. The system accommodates aggregate functions, multiple conditions in WHERE clause, join operations, advanced clauses like ORDER BY, GROUP BY and HAVING. The natural language statement goes through various stages of Natural Language Processing like morphological, lexical, syntactic and semantic analysis resulting in SQL query formation. Intelligent Interface is the need of database applications to enhance efficient interaction between user and DBMS. The research focuses on making the system more dynamic.Improvements have been introduced to the system by incorporating preprocessing of text, named entity recognition, building hierarchical relations, semantic similarity and negation handling using dependency graphs.
Natural language processing (NLP), SQL, semantic similarity, context, named entity recognition,dependency graphs
Research on Paper Intelligent Plagiarism Detection Method Based on Idea Tendency
, Hanmo Zhang1
, Wangyang Liu2
, Yang Cao2
and Weiping Fu1
College of Computer Science and Technology, Harbin Engineering University, Harbin, China and 2
CETC big data research institute company limited, China.
This paper proposes a method that analyzes the tendency of the paper for the originality detection of papers.This method is aiming at the behavior of idea plagiarism. The plagiarism detection task can be divided into three steps which are retrieval of candidate source document set, selection of source document set and post processing. The texts are modeled by the topic model and then are divided into several topic blocks by combining the topic modeling with the classical text segmentation method TextTiling. Then, the topic blocks which may be the source block of plagiarism are filtered out by calculating topic similarities among these topic fragments. On this basis, topic words of the topic block to be measured are extracted and its idea tendency aiming at each topic word are analyzed by the sentiment analysis method based on the dictionary. And further the source topic block set is filtered out based on the idea tendency Finally, a method of evaluating the original degree of the text is given for the intelligent plagiarism detection by the fusion of topic similarity and the result of the analysis of idea tendencies.Through experimental verification, this method can effectively evaluate the originality of the paper and has a good accuracy.
Plagiarism detection, Topic modeling, Text Segmentation, Idea Tendency
Automatic Text Summarization of Legal Cases: A Hybrid Approach
Varun Pandya, School of Computer Engineering, Pandit Deendayal Petroleum University, Gandhinagar, India
Manual summarization of large bodies of text involves a lot of human effort and time, especially in the legal domain. Automatic Text summarization is a constantly evolving field of Natural Language Processing(NLP) discipline of the broader Artificial Intelligence Field. Lawyers spend a lot of time preparing legal briefs of their client’s case files. In this paper a hybrid method for automatic text summarization of legal cases using k-means clustering technique and tf-idf(term frequency-inverse document frequency) word vectorizer is proposed. The summary generated by the proposed method is compared using ROGUE evaluation parameters with the case summary as prepared by the lawyer for appeal in court. Further, suggestions for improving the proposed method are also presented.
Automatic Text Summarization, Legal domain, k-means clustering, tf-idf word vectors
POS Tagger model for Kannada Text with CRF++ and Deep Learning approaches
M. Rajani Shree1
and Dr Shambhavi B R2
Assistance Professor, Department of Computer Science & Engineering, School of Engineering and Technology, Jain Deemed-to-be University, Bengaluru, India and 2
Associate Professor, Department of Information Science & Engineering BMSCE, Bengaluru, India
Computational Linguistics is one of the interesting topics in the research field of Computer Science. This paper presents training for Part of Speech (POS) tagging on Kannada words using two techniques. First approach is supervised machine learning technique CRF++0.50 (Conditional Random Field). The second approach is a combination of word embedding and deep learning techniques. The total dataset used for this implementation is 1200 tagged Kannada sentences downloaded from Technology Development for Indian Languages (TDIL). We divided the dataset into 1100 sentences (13,600 words) as training data and 100 sentences (1053 words) as test data. The BIS (Bureau of Indian Standards) tagset is used in this work in which 27 major POS tags have been considered. An accuracy obtained through CRF++0.50 tool is 76% and that with deep learning technique is 71%. The precision, recall and f-score of each tag using both techniques have been acceptable.
Part of Speech tagging, Conditional Random Field, Deep Learning, Word Embedding
Challenges in Rule Based Machine Translation from Marathi to English
Namrata G Kharate1
, Dr.Varsha H. Patil2
Department of Computer Engineering, VIIT,Pune, Maharashtra, India and 2
Head of Department, Department of Computer Engineering, MCOERC, Nashik, Maharashtra, India
Machine translation is being carried out by the researchers from quite a long time. However, it is still a dream to materialize flawless Machine Translator and the small numbers of researchers has focussed at translating Marathi Text to English. Perfect Machine Translation Systems have not yet been fullybuilt owing to the fact that languages differ syntactically as well as morphologically. Majority of the researchers have opted for Statistical Machine translation whereas in this paper we have addressed the challenges of Rule based Machine Translation. Thepaper describes the major divergences observed in language Marathi and English and manychallenges encountered while attempting to build machine translation system form Marathi to English using rule based approach.As there are exceptions to the rules and limit to the feasibility of maintaining knowledgebase, the practical machine translation from Marathi to English is a complex task.
NLP, Machine Translation, English, Marathi, grammar
Word Sense Disambiguation: Review
Moape Tebatso Gorgina1
and Mvelase Promise2
Department of Computer Science, University of South Africa, Florida Park, Roodepoort and 2
MediDeniz Software, Old Street, New York, USA
Natural Language Processing(NLP) applications are confronted with ambiguitiesand subsequent difficulties whenprocessing texts.These ambiguities in NLP are called Word Sense Disambiguation. Word sense disambiguation (WSD) is the process ofdisambiguating the sense of an ambiguous word when the wordhas more than one sense. To circumvent this problem,several approaches such knowledge based, supervised, unsupervised and semi-supervised machine learning algorithms have been proposed and used in different languages. In this paper, we review different approaches used to solve the problem of WSD and their effectiveness, a brief history of WSD, challenges and applications.
TGRID Location Service in Ad-Hoc Networks
Baktash Motlagh Farrokhlegha, Department of Computer, Technical & vocational University, Urmia, West Azarbayjan, Iran
Geographic addresses are essential in position-based routing algorithms in mobile ad hoc networks, i.e. a node that intends to send a packet to some target node, has to know the target's current position. A distributed location service is required to provide each node's position to the other network nodes. Hierarchical Location Service (HGRID) has been known as a promising location service approach. In this paper we present a new approach called TGRID and describe the performance of a
novel multi-level Tree-walk grid location management protocol for large scale ad hoc networks. The Treewalk grid location service mechanism is evaluated by GLOMOSIM against well known location service protocol HGRID when increasing node density and node speed. It is observed that TGRID outperforms HGRID in terms of packet delivery fraction and storage cost and also maintains low control overhead in a uniformly randomly distributed network.
Location based routing, location service, location management, Mobile Ad Hoc Networks, HGRID, and TGRID
Automated Regression Tests and Automated Test Optimisation for GETRV
Neil Kevin Patalita Arcolas and Shahid Ali, Department of Information Technology, AGI Institute, Auckland, New Zealand
Regression testing is a type of testing that is performed to validate that new changes pushed to the system does not have any adverse effect to the existing features. Automated regression testing greatly reduces the time spent by testers to perform these repetitive and mundane tests and allows them to work on more critical tests. The first problem addressed in this project is to add two automated regression scripts to increase test coverage of the existing test automation framework. The second problem is to optimise the automated regression test run to reduce the test run times. Additionally, to improve the automated test run times, redundant expressions were removed and handled in the outermost loop of the automated test run. The project resulted in the addition of two automated test scripts for the automated test run and a significant test run time reduction of at least 60%.
Automated regression, Agile scrum, Automated test run
Predicting Weather events using Soft Computing Techniques
Hatim Aljuaed and Mohamed Alghamdi, Department of Computer Science and Computer Engineer, Umm-Al Qura University, Makkah, Saudi Arabia
Weather Prediction has been widely needed for many applications with different purposes such as rainfall prediction, agriculture, pilot an aircraft and so on. It’s becomes so helpful in normal people’s lives to make more informed daily decisions especially in those who lives in countries with extreme or diverse weather conditions. Short term prediction is a critical mission as well to predict warnings that can help saving people’s lives and protect properties. This work suggests a system that is based on the integration between Fuzzy Logic (FL) and Artificial Neural Networks (ANNs). The usage of the two soft computing techniques enhances the accuracy nature of neural networks and the easy usage of the fuzzy logic. Upon this literature, the work will show how soft computing using fuzzy logic and neural networks solves the problem of weather forecast in Kingdom of Saudi Arabia, Makkah and extracting non-relative knowledge from the known facts from different types of data.
Neural Networks, Fuzzy logic, soft computing, weather prediction.
An Overview of Steganography - A Data Hiding Technique
Priya Pareek and N.Monica, Student, Department of Computer Science and Engineering GMRIT, Rajam
Steganography is a technique which is used to hide the secret data in the form of embedded messages, simply it is covered by the other messages. The word Steganography is derived from two ancient Greek word names "Stegano"(which means hiding or covering) and "graph"(meaning to write). Steganography is very different from cryptography. Cryptography is a technique which is used to create or generate the code which keeps information very secure. In simple comparisons, cryptography is the one that cannot be decrypted without the proper knowledge of encrypted data key whereas, steganography is the one which can be easily used to encrypt or decrypt the data
Steganography, Data hiding, Cryptography, Types of Steganography
Opportunistic touting based on Store carry and forward algotithm to avoid discontinuity in VANET
Mohamed Anis MASTOURI and Salem Hasnaoui, Communication systems – Sys’Com Laboratory National school of engineers of Tunis - ENIT Tunis, Tunisia
The publish/subscriber model according to DDS is essentially characterized by the decoupling between the participants and the interaction many-to-many. These features are desirable properties for constructing distributed applications in the context of a mobile environment such as VANET. Which are currently receiving in-creased attention from manufacturers and researchers to improve safety on the roads or help drivers. So the goal of this paper is to design a solution of routing in order to avoid discontinuity in such environments based on publish subscribe paradigm.
VANET, publish-subscribe, MANET, opportunistic, routing
A Group Decision Making for Coastal Wetland Park Selection Approach Using Topsis and Interval-Valued Fuzzy Numbers
and Ming-Ching Lee2
Department of Tourism, Food & Beverage Management, Chang Jung Christian University, 71101 Tainan, Taiwan
Department of Hospitality Management, Taiwan Shoufu University, 72153 Tainan, Taiwan
Due to the increasing awareness of environmental and social issues, sustainable coastal wetland park selection becomes an important problem. The aim of this paper is to develop a new performance evaluation method for multi-attribute decision making (MADM) problems in a group decision environment, based on combining an integrated group Techniques for Order Preferences by Similarity to Ideal Solution (IG-TOPSIS) and interval-valued fuzzy numbers. Coastal wetland park selection often involves uncertain information due to the subjective nature of human judgments. Because human beings are more suitable using linguistic terms rather than crisp values or precise numbers to express what they perceive, the rating values can be expressed in linguistic terms. These linguistic terms, however, are often imprecise or vague. The interval-valued fuzzy sets can provide more flexibility than ordinary fuzzy sets in representing vague or uncertain information. This paper presents an interval-valued fuzzy IG-TOPSIS method, which aims to solve MADM problems in which the preferences of different decision makers are considered and expressed clearly using the concepts of interval-valued fuzzy sets. A case study for evaluating the performances of several sites for coastal wetland park selection is conducted to illustrate the effectiveness of this method.
Interval-valued fuzzy sets; TOPSIS; Integrated group TOPSIS; MADM
Simulation of Steering a Self-Driving Car Using 1) PID Controller 2) Neural Network
Deeheem Ansari and Gurtej Kochar, Netaji Subhas Institute of Technology (NSIT), New Delhi, India
Over the last few year, self-driving vehicles have expanded dramatically. They look promising in decreasing traffic accidents and congestion on roads. In order for this approach to work, simulations tend to be a cheaper, more efficient, and a safer way than live testing. In this paper, we present a simulator that tests the self-driving car’s CNN model in a virtual environment. To avoid the hassle of collecting huge amount of training and testing data for preliminary testing, we also propose a method to collect data (images, steering value and throttle value) using PID controllers in the virtual environment itself.
Autonomous vehicle, Neural network, PID controller, Self-driving vehicle, Simulation.
EACO: An Enhanced ANT Colony Optimization Algorithm for Task Scheduling in Cloud Computing
and Richa Jain2
Banasthali Vidypith, Tonk, Rajasthan, India
Cloud Computing is emerging as an influential architecture to perform complex and large scale computing. It provides on-demand access to services on “Pay –as –you-go” method. In cloud computing environment Task Scheduling is an essential technique that is required for allocating tasks to appropriate resources for proper resource utilization and optimizing overall system performance. In this paper, the enhanced Ant Colony Optimization (EACO) algorithm has been proposed that serves improved task scheduling with minimum makespan while maintaining cost. This algorithm mainly contributes in minimizing total completion time for scheduling tasks on resources. This is attained by splitting the ordered submitted tasks into bunches -the sub list of tasks. The main goal of EACO is to minimize total execution time. The proposed algorithm EACO is simulated using CloudSim toolkit and compared with existing nature inspired algorithm. The experimental results show that presented algorithm improves result in terms of makespan
Cloud computing, task scheduling, ant colony optimization (ACO), makespan.
Mobile Cloud Computing in Healthcare Using Dynamic Cloudlets for Energy-Aware Consumption
Manoj Muniswamaiah and Dr. Charles Tappert, Seidenberg School of CSIS, Pace University, White Plains, New York
Mobile cloud computing (MCC) has increasingly been adopted in healthcare. The use of mobiles devices by healthcare professionals (HCPs) has transformed clinical practice. Increasing use of mobile devices has led to the growth in the development of medical software applications for these platforms. There are different applications which help HCPs with many important tasks. Mobile cloud computing has helped HCPs in better decision making and improve patient care. MCC enable users to acquire the benefits of cloud computing services to meet the healthcare demands. However, the restrictions posed by network bandwidth and mobile device capacity has brought challenges with respect to energy consumption and latency delay. In this paper we propose dynamic energy consumption mobile cloud computing model (DEMCCM) which addresses energy consumption by healthcare mobile devices by using dynamic cloudlets.
Mobile cloud computing, cloud computing, healthcare, cloudlets, dynamic programming.
Addressing Application Portability Challenge In cloud Computing
Jayaprakash Arjarapu, Oracle India Private Limited, Hyderabad, India
The cloud computing is a buzz word for all the technologists and it is the recent revolution in IT industry. Cloud computing drastically changing the way IT service providers operate and moved the ball into the court of customers to some extent. To the business community where customers were spending more time on the IT rather than their actual business, cloud computing is a wonderful opportunity which allows them to spend more time on their business than IT. As there is a huge demand in the market for the cloud application platforms, it resulted in a large number of platform offerings which are readily available in the market. Apart from the benefits customer can enjoy from this cloud computing, there are also different challenges which need to be addressed before customers fully adopt this cloud computing to their business. One of such challenge is application portability. In this research work, we are trying to bring out application portability challenge and focus on addressing this challenge of porting applications from one platform to other. Our research also opens the door to future research directions towards investigating the improvement of applications portability across different cloud platforms
Airport Cyber Security & Cyber Resilience Controls
Alex Mathew, Department of Computer Science & Cyber Security, Bethany College, USA
Cyber Security scares are the main areas of demerits associated with the advent and widespread of internet technology. While the internet has improved life and business processes, the levels of security threats have been increasing proportionally. As such, the web and the related cyber systems have exposed the world to the state of continuous vigilance because of the existential threats of attacks. Criminals are in the constant state of attempting cybersecurity defense of various infrastructures and businesses. Airports are some of the areas where cybersecurity means a lot of things. The reason for the criticality of cybersecurity in airports concerns the high integration of internet and computer systems in the operations of airports. This paper is about airport cybersecurity and resilience controls. At the start of the article is a comprehensive introduction that provides a preview of the entire content. In the paper, there are discussions of airport intelligence classification, cybersecurity malicious threats analysis, and research methodology. A concise conclusion marks the end of the article.
Cyber Security ,IOT, Resilience Controls