Nowadays, communication between patient and doctor during an appointment has changed significantly owning to the opportunity that medical portals provide. Whether or not necessarily appreciated by the doctors, the patients became more aware of the first symptoms’ suggesting a particular disease and the medical procedures that apply as a standard. Estimating amount of reliable factual medical information in a document is carried out by parametrizing space of digital documents and dividing it into subsequent layers that represent distribution of the system responses computed as random variables to a query about medical information. Analyzed are the following attributes: dynamism of decrease of query words numbers in the documents, precision, recall in the metric space layers, their mutual correlation and specifically the amount of reliable medical information in the documents. Sensitivity of estimators is explored in order to determine the final decision about further browsing digital documents of the metric space for more medical information that satisfies the user’s need. For identification of the true positive information in the space layer and then, in each document of this layer, matching of medical terminology with the document contents, is processed following binary Boolean search space model.
DevOps has become an emerging force within the Information Technology field in today's development/operations climate. Information security within a DevOps environment has become a focal point for most organizations that have implemented the DevOps methodology and its principles. In most cases, the ability to secure a DevOps environment and the organization's ability to adhere to, and comply with, industry specific standards, frameworks and best practice is an integral part of information security within a DevOps environment. This investigation aims to address those issues that may arise when an organization seeks to adhere to and comply with industry standards, frameworks and best practice in a manner that does not limit the velocity of the organization's automated delivery/deployment pipeline. This study investigates this by collecting and analyzing industry and academic literature; and through a prototype demonstration, understanding technical compliance and its requirements within a DevOps environment, using existing industry tools and solutions.
A huge amount of user generated content related to movies is created with the popularization of web 2.0. With these continues exponential growth of data, there is an inevitable need for recommender systems as people find it difficult to make informed and timely decisions. Movie recommendation systems assist users to find the next interest or the best recommendation. In this proposed approach the authors apply the relationship of user feature-scores derived from user-item interaction via ratings to optimize the prediction algorithm's input parameters used in the recommender system to improve the accuracy of predictions with less past user records. This addresses a major drawback in collaborative filtering, the cold start problem by showing an improvement of 8.4% compared to the base collaborative filtering algorithm. The user-feature generation and evaluation of the system is carried out using the `MovieLens 100k dataset'. The proposed system can be generalized to other domains as well.
The prevalence of social media has made information sharing possible across the globe. The downside, unfortunately, is the wide spread of misinformation. Methods applied in most previous rumor classifiers give an equal weight, or attention, to words in the microblog, and do not take the context beyond microblog contents into account; therefore, the accuracy becomes plateaued. In this research, we propose an ensemble neural architecture to detect rumor on Twitter. The architecture incorporates word attention and context from the author to enhance the classification performance. In particular, the word-level attention mechanism enables the architecture to put more emphasis on important words when constructing the text representation. To derive further context, microblog posts composed by individual authors are exploited since they can reflect style and characteristics in spreading information, which are significant cues to help classify whether the shared content is rumor or legitimate news. The experiment on the real-world Twitter dataset collected from two well-known rumor tracking websites demonstrates promising results.
In information retrieval, documents are usually retrieved using lexical matching which matches where words in a user's query with words found in a set of documents. A significant model used in information retrieval is the vector space model where these words are represented as a vector in space and are assigned weights using a favorite weighting technique called TFIDF (Term Frequency Inverse Document Frequency). In this thesis, we have devised three new weighting techniques to improve the TFIDF weighting technique. The first technique is Dispersed Words Weight Augmentation (DWWA) which gives more weight to the words distributed in most of the document's paragraphs; we consider that those words are more significant than words found in few paragraphs. The second technique is called Title Weight Augmentation (TWA) which gives more weight to the words found in the document's title and first paragraphs. The third technique is called First Ranked Words Weight Augmentation (FRWWA) which increments further the weight of the most frequent words in a document. We tested the three techniques, and we found more relevant documents were retrieved in our system.
Often, textual data is accompanied by images in an attempt to help the user better understand the content. Drawing inspiration from this, it would be interesting to transform any given text into a digital image collage, an artistic graphic object. To automate the process, we developed a software system that uses Natural Language Processing to summarize large paragraphs of text, retrieves images from databases, and arranges them with a Genetic Algorithm on a digital canvas. Given that certain images may be more important to the user, the software computes emphasis factors for scaling the images in the collage. We performed computational experiments to validate each module of our system and initiated user studies to examine their perception of the collages.
The volume of RDF data continues to grow over the past decade and many known RDF datasets have billions of triples. A grant challenge of managing this huge RDF data is how to access this big RDF data efficiently. A popular approach to addressing the problem is to build a full set of permutations of (S, P, O) indexes. Although this approach has shown to accelerate joins by orders of magnitude, the large space overhead limits the scalability of this approach and makes it heavyweight. In this paper, we present TripleBit +, a fast and compact system for updating RDF data. The design of TripleBit + has two salient features. First, the efficient maintenance strategies of TripleBit + reduces both the overhead to update data and indexes. Second, effective maintenance technologies to handle online updates over RDF repositories are proposed. Our experiments show that TripleBit + outperforms RDF-3X, MonetDB, BitMat on LUBM, UniProt, and BTC 2012 benchmark queries and it offers orders of mangnitude performance improvement for some complex join queries. Our design also yields high task rates as high as 660,000 per second and fast average response time of task which is faster than x-RDF-3X and PostgreSQL.
Devolved governments such as the county and regional governments around the world have a constitutional responsibility to find sustainable ways through which they can meet material, social, and economic responsibilities of improving the quality of the lives of their citizens by providing high-quality services and decent work for their employees. The 2014-2017 Kenya's Council of Governors strategic plan postulated enactment of a knowledge management strategy where good practices and lessons learnt within any county government should be documented and disseminated in appropriate forums to other counties. However, the 2017-2022 strategic plan indicates that there is lack of a structured mechanism for systematic knowledge sharing and organizational learning among the county governments despite the council's effort to share information through statutory annual reports, devolution conferences and quarterly sectoral committee meetings. But, the 2017-2022 strategic plan envisions a systematic mechanism for sharing experiences among the county governments. The intention of this study was to assess the current knowledge management practices among the county governments in Kenya; to identify, and articulate knowledge management concepts that are useful to the public services sector among devolved governments in developing countries; and to model these practices into a framework that can support continuous sharing of experiences, lessons and innovations within and among the county governments in Kenya. Theoretical frameworks and models of knowledge management in governance, governments and e-governments were considered and a conceptual framework for successful knowledge management initiatives among county and regional governments was formulated. The proposed conceptual framework was evaluated using a focus group discussion with participants drawn from the Council of Governors' Maarifa Center employees. The study proposes a framework to facilitate effective sharing of experiences among county employees, between different county governments and to manage and enhance knowledge management initiatives among the devolved governments. The study results indicate some sporadic nascent knowledge management practices rather than well planned initiatives within the counties. The study provides recommendations for the Council of Governors and other policy makers on how to manage knowledge management initiatives, while suggestions for future research directions for researchers with similar interests are given.
With the growth of heterogeneous data generation sources low-quality data volumes are expanding on a daily basis. This research proposed SimFiller: similarity-based missing (null) values filling algorithm, to enhance the quality of data for the data mining process. The proposed algorithm calculates the similarity of record pairs from the input data in such a way that at least one member of the pair has a non-null value for the attribute under consideration. After finding similar pairs, the algorithm fills the missing values by considering the pair having greatest similarity under the specified similarity threshold. The quality of resulted data is evaluated by analyzing the classification accuracy results for Audiology dataset. Five other missing values filling algorithms were selected and total six copies of filled Audiology dataset were created. All six copies of filled Audiology dataset were tested for their classification accuracy. Results show a huge boost in classification accuracy for the copy of the dataset filled with the proposed algorithm and indicate that the quality of the dataset is enhanced. The proposed algorithm can also be tested on other datasets for filling their missing (null) values and can also be extended to remove other inconsistencies from the datasets.
This paper presents our smart parking solution implemented at “Lucian Blaga” University of Sibiu (LBUS) Romania, which consists in a hardware / software embedded system for managing the institution's parking lots, namely sharing the parking places in excess for people who are in traffic in neighbourhood of LBUS and are looking to park. Our solution is flexible, universal, applicable to all faculties that have car parks from Romania and not only, in university cities where the crowd is bigger, but also to other private institutions that own parking spaces inefficiently exploited. The advantages introduced are primarily economic, then social and even environmental. The first benefit is economic - for institutions which exploit their free parking spaces during the year, at different moments of time (holiday, afternoon, or time when people left from work) offering for people who are searching for. From a social point of view, ensuring the convenience of drivers, reducing crowding especially at rush hours and reducing the time spent in search of a parking space, is also an important advantage. Reducing the cost of fuel required by cars that are looking for a parking space contributes to reducing pollution and improving air quality, having a positive impact on the environment.
The digital transformation of our society is happening. In this paper, we try to provide means to deal with this phenomenon. We introduce and examine a new approach to show how to capture the impact of digital transformation methodically, and by doing so, how to guide the complex unpredictable process of digitalization in our social environment. After showing related work on artifacts, on the representation of things, on modeling, and finally on models as artifacts, we present our new model-based approach, the flow of models we developed, namely models for object characterization, hypothetical story, prediction, and test/experiment/evaluation. Furthermore, we show the context of our research, the role of models in design, and how we broaden our research context from design to digital transformation. Before we conclude our paper, we illustrate our approach on an example from health care, in the scope of an international research project.
Blended learning is widely adopted by education agencies and organizations, as it is a flexible model in which face-to-face classroom practices are combined with computer-mediated activities. To overcome the limits of the loss of interaction between teacher and students and among students in distance learning, researchers proposed several solutions, conducting experiments in several teaching areas. Our interest is aimed at studying blended learning with a specific focus on those courses involving problem solving activities, through collaboration among students.Modern Learning Management Systems (LMS) allow to define virtual classrooms and offer various functionalities to support the class. At the same time, they are not designed to fully support all type of activities. Thus, they provide the possibility of integrating other more useful systems for more specific activities. A standard LMS has to be integrated using specific tools when problem solving activities are planned, to ensure effective collaboration among students. In this regard, there is no convergence towards a specific tool that can be used to carry out problem solving activities in collaboration.This paper aims to propose a minimal set of requirements for interaction mechanisms to support problem solving activities in a collaborative environment, in order to obtain better quality artifacts. We also report the results of a three-month experimental course (12 weeks) entitled ”Project Management: a look ahead”, based on blended learning and problem solving activities. The minimal set of requirements for interaction mechanisms was implemented using GitHub, that is not a teaching software, but it is a global software development tool which has powerful communication mechanisms. The results show that the aid of the proposed minimal set of requirements for interaction mechanisms significantly improves the quality of artifacts when problem solving activities are carried out.
Autonomous vehicles will be an integral part of future transportation systems, and the maritime industry is working towards developing methods to ensure safe autonomous ship operations. One of the major challenges in realizing autonomous ships is ensuring effective collision avoidance technologies. Autonomous vessels must have a higher degree of situation awareness to detect other vessels, predict their future intentions, and evaluate the respective collision risk. One step in achieving this goal is to predict other vessel trajectories accurately. In this paper, a data-driven approach to vessel trajectory prediction for time horizons of 5-30 minutes utilizing historical AIS data is evaluated. A clustering based Single Point Neighbor Search Method is investigated along with a novel Multiple Trajectory Extraction Method. Predictions have been conducted using these methods and compared with the Constant Velocity Method. Additionally, the Multiple Trajectory Extraction Method is utilized to evaluate estimated ship routes.
A patent examiner needs domain-specific knowledge to classify a patent application according to its field of invention. Standardized classification schemes help to compare a patent application to previously granted patents and thereby check its novelty. Due to the large volume of patents, automatic patent classification would be highly beneficial to patent offices and other stakeholders in the patent domain. However, a challenge for the automation of this costly manual task is the patent-specific language use. To facilitate this task, we present domain-specific pre-trained word embeddings for the patent domain. We trained our model on a very large dataset of more than 5 million patents to learn the language use in this domain. We evaluated the quality of the resulting embeddings in the context of patent classification. To this end, we propose a deep learning approach based on gated recurrent units for automatic patent classification built on the trained word embeddings. Experiments on a standardized evaluation dataset show that our approach increases average precision for patent classification by 17 percent compared to state-of-the-art approaches.
Analyzing data by visualization can help medical institutions make more informed decisions on the admission of future patients. A dataset provided by the University of California, Irvine, (UCI) Machine Learning Repository contains information on patients with diabetes that represents 10 years (1999-2008) of clinical care at 130 US hospitals. Charts on the diabetes dataset were created using Tableau, a data visualization software. In addition, an extensive analysis of the dataset was completed by providing possible reasons for the output found in the charts. The knowledge obtained from analyzing the diabetes dataset will help provide important information to medical institutions concerning future diabetic patients in need of hospital services.
We present a model to implement digital twins in sustainable agriculture. Our two-year research project follows the design science research paradigm, aiming at the joint creation of physical and digital layers of IoT-enabled structures for vertical farming. The proposed model deploys IoT to (1) improve productivity, (2) allow self-configuration to environmental changes, (3) promote energy saving, (4) ensure self-protection with continuous structural monitoring, and (5) reach self-optimization learning from multiple data sources. Our model shows how digital twins can contribute to the agrofood lifecycle of planning, operation, monitoring, and optimization. Moreover, it clarifies the interconnections between goals, tasks, and resources of IoT-enabled structures for sustainable agriculture, which is one of the biggest human challenges of this century.
Internet security and privacy is a trending issue. Different tools are being employed by different users to stay anonymous and safe. The research focused on impact of Big Five Inventory personality traits on use of VPN and proxy servers. The research was conducted in the Czech Republic. The respondents were 478 university students. Gender, age, and type of student's job were used as control variables. With regards to the results, conscientiousness, openness to experience, gender, and type of student's job influence the preference. It is less conscientious, more open to experience, male, and working full time in the field they study that use VPN and proxy servers.
Machine Learning is often associated with predictive analytics, for example with the prediction of buying and termination behavior, with maintenance times or the lifespan of parts, tools or products. However, Machine Learning can also serve other purposes such as identifying potential errors in a mission-critical large-scale IT process of the public sector. A delay of troubleshooting can be expensive depending on the error's severity- a hotfix may become essential. This paper examines an approach, which is particularly suitable for Static Code Analysis in such a critical environment. For this, we utilize a specially developed Machine Learning based approach including a prototype that finds hidden potential for failure that classical Static Code Analysis does not detect.
Digital marketing sector is an expanding sector with increasing number of customers, product and service types. Dynamic structure of websites, rapid changes in stock levels in e-commerce websites and dependency of existing systems to humans causes digital marketing agencies to require autonomous systems for customer account management. So the aim of this study is developing an autonomous intelligent system that operates integrated with Google Ads platform. The developed system makes optimization in return-on-investment, conversions, advertising texts and profit using Ackermann Feedback Control Algorithm and State Transition Matrices.
This article discusses clustering approach to recommender systems acceleration and presents application of multi-clustering algorithms in the recommender systems based on collaborative filtering. It is explained the motivation for multi-clustering usage in comparison to clustering techniques, as well as results of experiments. Multi-clustering is variously defined in literature, however the common issue is its multiple views of one dataset. Different views may represent distinct aspects of the same data, adapting the most appropriate one to the current problem. In recommender systems domain it can be applied as a tool for precise modelling neighbourhood of object the recommendations are generated to. This article presents results of experiments demonstrating multi-clustering advantage over traditional clustering in neighbourhood determination.
