The Internet of Things (IoT) primary objective is to make a hyper-connected world for various application domains. However, IoT suffers from a lack of interoperability leading to a substantial threat to the predicted economic value. Schema.org provides semantic interoperability to structure heterogeneous data on the Web. An extension of this vocabulary for the IoT domain (iot.schema.org) is an ongoing research effort to address semantic interoperability for the Web of Things (WoT). To design this vocabulary, a central challenge is to identify the main topics (concepts and properties) automatically from existing knowledge in IoT applications. We designed knowledge extraction for the WoT (KE4WoT) to automatically identify the most important topics from literature ontologies of three different IoT application domains: 1) smart home; 2) smart city; and 3) smart weather-based on our corpus consisting of 4500 full-text conference and journal articles to utilize domain-specific knowledge encoded within IoT publications. Despite the importance of automatically identifying the relevant topics for iot.schema.org, up to know there is no study dealing with this issue. To evaluate the extracted topics, we compare the descriptiveness of these topics for the ten most popular ontologies in the three domains with empirical evaluations of 23 domain experts. The results illustrate that the identified main topics of IoT ontologies can be used to sufficiently describe existing ontologies as keywords.

Auto-Extraction and Integration of Metrics forWeb User Interfaces

Training data for user behavior models that predict subjective dimensions of visual perception are often too scarce for deep learning methods to be applicable. With the typical datasets in HCI limited to thousands or even hundreds of records, feature-based approaches are still widely used in visual analysis of graphical user interfaces (UIs). In our paper, we benchmarked the predictive accuracy of the two types of neural network (NN) models, and explored the effects of the number of features, and the dataset volume. To this end, we used two datasets that comprised over 4000 webpage screenshots, assessed by 233 subjects per the subjective dimensions of Complexity, Aesthetics and Orderliness. With the experimental data, we constructed and trained 1908 models. The feature-based NNs demonstrated 16.2%-better mean squared error (MSE) than the convolutional NNs (a modified GoogLeNet architecture); however, the CNNs’ accuracy improved with the larger dataset volume, whereas the ANNs’ did not: therefore, provided that the effect of more data on the models’ error improvement is linear, the CNNs should become superior at dataset sizes over 3000 UIs. Unexpectedly, adding more features to the NN models caused the MSE to somehow increase by 1.23%: although the difference was not significant, this confirmed the importance of careful feature engineering.

The data binding pattern is an established technique to couple user interface (UI) elements and data objects. Various markup languages (e.g. Microsoft XAML, Adobe MXML) integrate advanced data binding concepts in order to ease application development. However, the HTML standard does not embrace means for data binding although being the Web markup language supported by millions of Web programmers. Therefore, we propose a standard-compliant WebSocket-based Data Binding (WebSoDa) framework. The WebSoDa framework synchronizes data objects and UI elements by orchestrating a Microdata-based data binding language as well as a client-side and a server-side messaging component. Thus, developers may speed up the tedious task of implementing binding associations in Web applications.