Evaluation Metrics for Health Chatbots: A Delphi Study

Background In recent years, an increasing number of health chatbots has been published in app stores and described in research literature. Given the sensitive data they are processing and the care settings for which they are developed, evaluation is essential to avoid harm to users. However, evaluations of those systems are reported inconsistently and without using a standardized set of evaluation metrics. Missing standards in health chatbot evaluation prevent comparisons of systems, and this may hamper acceptability since their reliability is unclear. Objectives The objective of this paper is to make an important step toward developing a health-specific chatbot evaluation framework by finding consensus on relevant metrics. Methods We used an adapted Delphi study design to verify and select potential metrics that we retrieved initially from a scoping review. We invited researchers, health professionals, and health informaticians to score each metric for inclusion in the final evaluation framework, over three survey rounds. We distinguished metrics scored relevant with high, moderate, and low consensus. The initial set of metrics comprised 26 metrics (categorized as global metrics, metrics related to response generation, response understanding and aesthetics). Results Twenty-eight experts joined the first round and 22 (75%) persisted to the third round. Twenty-four metrics achieved high consensus and three metrics achieved moderate consensus. The core set for our framework comprises mainly global metrics (e.g., ease of use, security content accuracy), metrics related to response generation (e.g., appropriateness of responses), and related to response understanding. Metrics on aesthetics (font type and size, color) are less well agreed upon—only moderate or low consensus was achieved for those metrics. Conclusion The results indicate that experts largely agree on metrics and that the consensus set is broad. This implies that health chatbot evaluation must be multifaceted to ensure acceptability.

Reading Speed Performance based on Interface Design Elements for Web Interfaces

The development of technology and internet has contributed to more people viewing on-screen content and an increase the amount of material that can be read directly from a Web-based platform. The transformation from paper-based materials to digital materials on the Web, more or less has affected the readability of the materials in which information presented on the printed and web-based documents is dissimilar. However, only few studies have been carried out on the presentation of information specifically, the readability of the text in web-based platform. Meanwhile, readability is one of the crucial aspect of Web usability and could affect the success of a website. The speed of reading is regarded as a good measure of readability. Therefore, this study aims to measure the readers reading speed for web interfaces. This study followed a quantitative analysis method that used experiments to collect data and involved university students as the test subjects. A Web-based Reading Speed Test System (RSTS) was developed for data collection and analysis purpose. The findings of this study found that the web interface with the combination of blue font color, yellow background color and Arial font type have the highest reading speed. In contrary, web interface that consists of combination black font color, white background color and Arial font type have the lowest reading speed. In addition, the reading speed of male participants was slightly higher than the female participants. To conclude, the results of this study signify that the interface design had a significant impact on the students reading speed for web pages.

Problems of Attributing the Publications of the Old Believer Ovchinnikov Printer

This article is dedicated to the problem of attributing the publications of illegal Old Believer printers in the second half of the nineteenth century: such publications do not have bibliographical data or have false information about the place of publication. At the centre of the author’s attention is the largest and longest-running printing company in the Old Believer market that belonged to the brothers Ovchinnikov. For publications from the eighteenth to the early nineteenth centuries, the main element which allows us to attribute a book to this or that printer is decoration; however, this criterion is insufficient for books from the second half of the nineteenth century. Old Believer merchants were publishers of books in the eighteenth and early nineteenth centuries; they placed their orders in rather large printing companies with an established set of decorations. Even in the printers they started at the end of the eighteenth century, they followed the tradition of using individual and distinctive ornamentation. In the second half of the nineteenth century, peasants or townsmen were publishers of books. Lacking sufficient funds and due to the illegal character of their activities, their printers were small and mobile, and they printed books on cheap paper in order to lower pro-duction costs. With only rare exceptions, the decoration of such books was not rich: either it was absent in general or there were only typeset decorations. If they did have headpieces, different printers used the same models. As such, ornamentation cannot be used as a universal indicator for attributing these publications, although it can be part of a set of such indicators. Fonts play the main role in defining an edition, but it is neces-sary to create a reference guide of fonts if this indicator is to be used practically. This article reflects the results of an analysis of fonts and decorative elements in publications from the Ovchinnikov printer. Very many such publications used a great variety of fonts simultaneously. According to indirect data, the books date from the 1860s. They might initially be taken for sammelbände since the type of paper changes along with the fonts. However, the integrity of the text during the transition from one font type to the next undermines this assumption. The article poses the question of how to explain the peculiarities of such publications. There is no one-dimensional answer. However, the author makes some suggestions: it is possible that at some point the printer was dispersed, with several printing presses placed in different houses. When a comparatively large volume was produced, the material was shared out and printed on these separate press-es. However, this hypothesis provokes the question: why do blocks of text printed in one font not follow each other in sequence? The article contains examples of the printer’s fonts and artistic elements the author identified.

Paralinguistic Tools in an Electronic Literary Text: Types, Structure, Semantic and Pragmatic Characteristics

The rapidly growing trend of information visualization in the modern communication space determines the relevance of this research, which is based on the analysis of the electronic literary text non-verbal components and their visual and graphic design. The purpose of the article is to identify the main methods of verbal components of graphic selection in electronic literary texts and their functional analysis. The material is represented by electronic literary texts based on Adobe Flash and HTML. The main methods of the research are general scientific methods such as monitoring and description in conjunction with the method of comparative linguistic analysis. The types of graphic highlighting that are traditionally used in written/printed text are also reflected in electronic literary texts. They include: 1. Graphic text segmentation; 2. Kerning (increasing the interval between letters); 3. Line spacing; 4. Font variation (changes in font type and color); 5. Capitalization (using uppercase letters instead of lowercase); 6. Italics, underscores, optional quotation marks (and various combinations of these techniques); 7. Inclusion of other sign systems elements (Latin/ Cyrillic/ Runic letters, numbers, encoding elements and etc.). As a result, the author concludes that techniques for verbal components graphic selection in electronic literary texts have several areas of functional load. Paragraph-based tools intensify the implementation of emotional and symbolic functions associated with the impact on the emotional background of the electronic literary text perception and the expression of its concepts.

Louder = Larger = Clearer

Past research found robust metamemory illusions about the effects of font type, word-pair identity, volume, and font size on memory that are assumed to share a common cause, such as fluency. The current study simultaneously assessed all four metamemory illusions from vignettes alongside items assessing the belief that fluency benefits memory and that more is generally better. The typical metamemory illusions replicated in all samples. Confirmatory factor and structural equation modeling confirmed that at least the perceptual metamemory illusions (font type, volume, font size) can be explained by one latent factor, which was, however, not related to latent factors capturing the belief that fluency benefits memory or that more is better.

E-LEARNING: APPROPRIATE E-MATERIALS FORMATTING FOR USERS HEALTH

E-learning has a direct relationship with public health as e-materials are making a huge amount of near workload for e-learners' eyes. As it is known, the huge near workload is one of the main reasons for the myopia development of nowadays population. The visual system can quickly become overloaded, especially by inappropriate e-material formatting. Based on theoretical research, incorrect formatting is used in most e-materials based on wrong recommendations. Recommendations and methodologies are not up-to-date for screen use. According to publications in the period of more than 20 years, near work and accommodation are the key factors for myopia development and progression. Appropriate formatting parameters of e-materials play an important role in reducing possible risk factors for myopia development. It could be achieved by using appropriate formatting parameters for e-learners. E-material font type must be perceptible and comprehensible from the screen, font size and line spacing must be appropriate for the reader based at least on its age and intellectual level, colour of background and text must help perception and reading process and could be different for each individual. All variables are important for individual and public health goals. New recommendations and automatization of the formatting process are developed to reach improvement.

License Plate Character Recognition using Riesz Fractional and Convolutional Neural Network

Automatic license plate recognition system is mostly used for identification of vehicles. This system is used in traffic monitoring, parking management and identification of theft vehicles. As in India the license plate regulations are not strictly followed, it is often difficult to identify the plate with different font type and character size. One more major problem in license plate recognition is low quality of images which affected via severe illumination condition. In this paper, a Riesz fractional mathematical model is proposed for enhancing the edges, which results in improving the performance of text recognition. The text in the license plate is recognized using the convolution neural network and the results showed better accuracy.

A Solution for Line Segmentation Problems in Sindhi Character Recognition System

The Sindhi language uses extended and modified set of the Arabic and Persian alphabets. It is the largest extension of the Arabic alphabet. Thus, Arabic or any other Arabic script based language character recognition system cannot recognize all characters of Sindhi. From character recognition point of view, Sindhi is a tough script. Sindhi’s cursive, context-sensitive characteristics, a large set of characters, highly similar shapes of the basic character, font-type variations, and size variations create high challenges for Sindhi character recognition research. In addition, line segmentation is a hard task as we have non-uniformity in the line heights. In this paper, we present an algorithm for segmenting the Sindhi text image into lines. The proposed algorithm solves the over-segmentation and under-segmentation problems in the line segmentation for Sindhi documents. The algorithm is tested on 100 text images of different Sindhi books and it has successfully segmented 99.95% lines correctly.

The Effects of Font Type on Reading Accuracy and Fluency in Japanese Children with Developmental Dyslexia

We administered rapid reading tasks in Japanese children (32 with typical development and 24 with developmental dyslexia), and investigated the effects of two different font types: Rounded-Gothic and Mincho style font. In the experiment, we used four kinds of stimuli: two scripts (paragraphs and random kana character non-words) in two font types (Rounded-Gothic and Mincho style font). In this experiment, the duration time, the number of errors and the number of self-corrections were measured during reading. Participants were asked which font type was easier to read. There was no significant difference in the duration time, the number of errors and the number of self-corrections between two types of fonts among the 56 participants. On the other hand, regarding subjective readability, the developmental dyslexia group reported that the Rounded-Gothic font was easier to read. There was a difference between objective and subjective readability. In this study, there was no difference in reading performance of Rounded-Gothic and Mincho style fonts in Japanese children with developmental dyslexia.