Search

In the era of Big Data there are a lot of new challenges – understanding, processing, and securing the data, assuring data quality, dealing with data growth and other challenges. One of the challenges is to identify and classify data sets in different systems which must follow the conditions defined by different regulations. The classification of these data sets can be automated using machine learning methods. The aim of the research is to provide machine learning methods for classifying sensitive data. The research is based on analysis and comparison of European Union legislation and scientific literature, which addresses issues of data classification using machine learning methods. Special attention is paid to sensitive data defined by the General Data Protection Regulation (GDPR). The main focus in this research is on supervised learning algorithms, where one of the most effective is Naïve Bayes classifier. In order to achieve good results, there is a need to find a proper training data set. Usage of hybrid methods provides a new way for increasing performance of classifiers. [ABSTRACT FROM AUTHOR]

Deep learning algorithms are becoming default solution for application in business processes where recognition, identification and automated learning are involved. For human identification, analysis of various features can be applied. Face feature analysis is most popular method for identification of person in various stages of life, including children and infants. The aim of this research was to propose deep learning solution for long-term identification of children in educational institutions. Previously proposed conceptual model for long-term re-identification was enhanced. The enhancements include processing of unexpected persons’ scenarios, knowledge base improvements based on results of supervised and unsupervised learning, implementation of video surveillance zones within educational institutions and object tracking results’ data chaining between multiple logical processes. Object tracking results are the solution we found for long-term identification realization. [ABSTRACT FROM AUTHOR]

Since INSPIRE directive was published in 2007, many datasets became publicly available in interoperable way. This makes possible the developing of wide range of applications (including web and mobile) that use this information as data source. In this article authors test all for download available datasets registered in Latvian spatial metadata catalogue and describes the technical details such as the feature type, feature count in dataset, geometry type, default coordinate reference system, standard name and derived parameters such as count of features in one hectare. Authors also provide qualitative assessment of web feature services and datasets registered in Latvian spatial metadata catalogue. Assessment is based on the results of validating this datasets/download services in INSPIRE geoportal metadata validator and using ETF test projects from github to test services compliance to INSPIRE Technical Guidance requirements. In results section short summary about all key parameters of datasets/download services are provided. This allows feature researchers and developers better understand technical characteristics of datasets and download services registered in Latvian spatial metadata catalogue. Moreover, this helps to choose the right tools to access information in most effective ways. [ABSTRACT FROM AUTHOR]

Since Infrastructure for Spatial Information in the European Community (INSPIRE) directive was published, Web Feature Services (WFS) more often are used to provide data exchange between information systems. WFS allows to receive spatial features as Geography Markup Language (GML), what is machine readable, but not the most effective way to disseminate information over World Wide Web (WWW). In this article authors research techniques to split WFS requests in most effective way and find the best solutions for downloading big datasets. For retrieval of big data sets best practice is to split request in many parts, sending each request only for a portion of information. Spatial and attributive filtering techniques can be applied for this splitting. The main findings of the study is that despite the server side limitations on maximum allowed through WFS downloaded portion size, it is possible to download the whole big (greater than 1000000 elements) datasets splitting request in many parts. This article summarizes pros and cons for attributive and spatial filtering techniques. The summarized pros and cons for WFS filtering can help to choose best practice solutions for information exchange in Spatial Data Infrastructure projects. [ABSTRACT FROM AUTHOR]

Historically, most web designers and their clients have approached the desktop side of any project first, while leaving the mobile part as a secondary goal that gets accomplished later. Lately there is a growing trend in the industry to flip this workflow and to begin with mobile considerations and then work up to a larger desktop version. Testing mobile versions of the websites is becoming an important task for developers and clients. Therefore, industries that aim their businesses to youth must take into consideration the quality of mobile versions of their websites. One of such industries is education. School websites are already published, it is necessary to test the usability of mobile devices at schools. From the point of usability, navigation in mobile versions of websites is an important factor. If the navigation structure and information display are not usable, it can greatly impact the overall usability of websites. This paper focuses on application of navigation testing methods using a head camera and video analysis to evaluate the mobile versions of websites. As the research object six Jelgava city secondary school websites are used. GoPro camera was chosen to take videos from the user's view point and to follow up activities and record comments. As the testing device iPad 2 tablet computer with the Internet browser Safari was selected. Before testing was executed, the participants needed to complete a questionnaire. Participants' behavior observation was performed and it shows a similar action pattern. The users' comments were used to complete the testing results. [ABSTRACT FROM AUTHOR]

Microclimate in classrooms can affect students' productivity. Since nowadays students and existing specialists of various fields, including engineering, most of the day spend indoors, it is important to ensure quality environment that would increase their productivity. It is important to pay attention to room microclimate factors that include temperature, air humidity and quality. Microclimate requirements are also regulated in the country legislation and organizations have to follow certain requirements. Therefore, it is important to perform detailed monitoring of microclimate. As technologies evolve there are various sensors offered in market that perceive certain microclimate parameters. These parameters inadequacy to recommended indicators can cause all kinds of health and work quality affecting problems as headaches, sleepiness, tiredness that results with lower concentration powers. In this research technological solutions for microclimate monitoring were introduced as a system prototype that measures microclimate parameters in classrooms. The monitoring system is based on the Arduino platform with sensors for measuring air temperature, humidity and carbon dioxide level. Additionally, a desktop application was developed to represent data remotely and receive warnings about microclimate indicator deviation from the recommended values. The developed system is used to monitor microclimate in the computer room at the Faculty of Information Technologies, Latvia University of Agriculture. During the research it was concluded that in non-ventilated room the CO2 level was always above the recommended value (1000 ppm) about 1500 ppm, temperature was slightly over the maximum recommended value (25 °C) but relative humidity was slightly over the minimum recommended value (30 %). The developed prototype can be scalable and adjusted for application in multiple buildings and organizations. [ABSTRACT FROM AUTHOR]

There are solutions in the market that enable brand holders to combine their serviceswithother suppliers of goods or services. However, there is no unified technological solutionavailable for e-commerce purchasing object management. Purchasing objects are objects that are generated automatically or manually createdas a result of the transaction, for example, a bus ticket, travel insurance payment, concert ticket orthird party liability insurance. Management of purchasing objects is an important e-commerce process for all involved parties. Now a variety of payment methods and information systems are available in the market that allows generating purchasing objects. E-tickets are an example of the purchasing objects, which are rapidly being introduced for use. Electronic ticket management is increasingly entrusted to mobile devices and applications. Market analysis shows that e-ticket management becomes more complicated because it is related to different systems, applications and interfaces in the transport sector. This study analyses the mobile ticket life-cycle management problems that are being solved in the development of the service providers' mutual integration solution architecture model. The unified public transport service scenario was developed, which consists of purchasing object generation, storing and transporting to service providers, transaction processors and customers. The developed purchasing object management model provides customers with opportunity to manage the financial transaction documents, while service providers will facilitate the document turnover and improve the business process. The proposed model has been piloted on a joint stock company micro-payment system infrastructure, integrating Latvian Railway external services. The study was conducted in the framework of Latvia, and is aimed to improve e-commerce solutions in Latvia. [ABSTRACT FROM AUTHOR]