Abstract Machine learning (ML) models are nowadays used in complex applications various domains, such as medicine, bioinformatics, and other sciences. Due to their black box nature, however, it may sometimes be hard understand trust the results they provide. This has increased demand for reliable visualization tools related enhancing ML models, which become a prominent topic of research community over past decades. To provide an overview present frontiers current on topic, we...

Text visualization has become a growing and increasingly important subfield of information visualization. Thus, it is getting harder for researchers to look related work with specific tasks or visual metaphors in mind. In this paper, we present an interactive survey text techniques that can be used the purposes search work, introduction gaining insight into research trends. We describe taxonomy categorization compare approaches employed several other surveys. Finally, results analyses...

Abstract Visualization of sentiments and opinions extracted from or annotated in texts has become a prominent topic research over the last decade. From basic pie bar charts used to illustrate customer reviews extensive visual analytics systems involving novel representations, sentiment visualization techniques have evolved deal with complex multidimensional data sets, including temporal, relational geospatial aspects. This contribution presents survey based on detailed categorization. We...

In machine learning (ML), ensemble methods such as bagging, boosting, and stacking are widely-established approaches that regularly achieve top-notch predictive performance. Stacking (also called "stacked generalization") is an method combines heterogeneous base models, arranged in at least one layer, then employs another metamodel to summarize the predictions of those models. Although it may be a highly-effective approach for increasing performance ML, generating stack models from scratch...

Abstract Over the past years, an increasing number of publications in information visualization, especially within field visual analytics, have mentioned term “embedding” when describing computational approach. Within this context, embeddings are usually (relatively) low‐dimensional, distributed representations various data types (such as texts or graphs), and since they proven to be extremely useful for a variety analysis tasks across disciplines fields, become widely used. Existing...

The sensemaking process of large sets text documents is highly challenging for tasks such as obtaining a comprehensive overview or keeping up with the most important trends and topics. Even though several established methods condensation summarization corpora exist, many them lack ability to account difference in prevalence between identified topics, which turn impedes quantitative analysis. In this paper, we therefore propose novel prevalence-aware method topic extraction, show how it can...

As the levels of automation and reliance on modern artificial intelligence (AI) approaches increase across multiple industries, importance human-centered perspective becomes more evident. Various actors in such industrial applications, including equipment operators decision makers, have their needs preferences that often do not align with decisions produced by black-box models, potentially leading to mistrust wasted productivity gain opportunities. In this paper, we examine these issues...

The machine learning (ML) life cycle involves a series of iterative steps, from the effective gathering and preparation data, including complex feature engineering processes, to presentation improvement results, with various algorithms choose in every step. Feature particular can be very beneficial for ML, leading numerous improvements such as boosting predictive decreasing computational times, reducing excessive noise, increasing transparency behind decisions taken during training. Despite...

Abstract The aim of this study is to explore the possibility identifying speaker stance in discourse, provide an analytical resource for it and evaluation level agreement across speakers. We also what extent language users agree about kind stances are expressed natural use or whether their interpretations diverge. In order perform task, a comprehensive cognitive-functional framework ten categories was developed based on previous work literature. A corpus opinionated texts compiled, Brexit...

Data visualization is of increasing importance in the Biosciences. During past 15 years, a great number novel methods and tools for biological data have been developed published various journals conference proceedings. As consequence, keeping an overview state-of-the-art research has become increasingly challenging both biology researchers researchers. To address this challenge, we reviewed especially performed Biosciences created interactive web-based tool, BioVis Explorer. Explorer allows...

Visualization for explainable and trustworthy machine learning remains one of the most important heavily researched fields within information visualization visual analytics with various application domains, such as medicine, finance, bioinformatics. After our 2020 state-of-the-art report comprising 200 techniques, we have persistently collected peer-reviewed articles describing categorized them based on previously established categorization schema consisting 119 categories, provided...

The automatic detection and classification of stance (e.g., certainty or agreement) in text data using natural language processing machine-learning methods creates an opportunity to gain insight into the speakers’ attitudes toward their own other people’s utterances. However, identifying presents many challenges related training collection classifier training. To facilitate entire process a classifier, we propose visual analytics approach, called ALVA, for annotation visualization. ALVA’s...

Online social media are a perfect text source for stance analysis. Stance in human communication is concerned with speaker attitudes, beliefs, feelings and opinions. Expressions of associated the speakers' view what they talking about up discussion negotiation intersubjective exchange. Taking thus crucial construction meaning. Increased knowledge can be useful many application fields such as business intelligence, security analytics, or monitoring. In order to process large amounts data...

Word Rain is a development of the classic word cloud. It addresses some limitations clouds, in particular lack semantically motivated positioning words, and use font size as sole indicator prominence. uses semantic information encoded distributional semantics-based language model – reduced into one dimension to position words along x-axis. Thereby, horizontal reflects similarity. Font still used signal prominence, but this supplemented with bar chart, well on y-axis. We exemplify by three...

Abstract Text visualization and visual text analytics methods have been successfully applied for various tasks related to the analysis of individual documents large document collections such as summarization main topics or identification events in discourse. Visualization sentiments emotions detected textual data has also become an important topic interest, especially with regard originating from social media. Despite growing interest this topic, research problem detecting visualizing...

Stacked generalization (also called stacking) is an ensemble method in machine learning that uses a metamodel to combine the predictive results of heterogeneous base models arranged at least one layer. K-fold cross-validation employed various stages training this method. Nonetheless, another validation strategy try out several splits data leading different train and test sets for then use only latter metamodel—this known as blending. In work, we present modification existing visual analytics...

Both the metadata and textual contents of scientific publications can provide us with insights about development current state corresponding community. In this short paper, we take a look at proceedings VINCI from previous years conduct several types analyses. We summarize yearly statistics different publications, identify overall authorship most prominent contributors, analyze community structure co-authorship network. also apply topic modeling to topics discussed in publications. hope that...

Comparing text documents is an essential task for a variety of applications within diverse research fields, and several different methods have been developed this. However, calculating similarity ambiguous context-dependent task, so many open challenges still exist. In this paper, we present novel method calculations based on the combination embedding technology ensemble methods. By using embeddings, instead only one, show that it possible to achieve higher quality, which in turn key factor...

Extreme weather events, such as flooding, are expected to increase in frequency and intensity. Therefore, the prediction of extreme assessment their local impacts urban environments, implementation adaptation measures becoming high-priority challenges for local, regional, national agencies authorities. To manage these challenges, access accurate warnings information about occurrence, extent, events crucial. As a result, addition official sources monitoring, citizen volunteered geographic...

Excitement or arousal is one of the main emotional dimensions that affects our lives on a daily basis. We win tennis match, watch great movie, get into an argument with colleague—all these are instances when most us experience excitement, yet we do not pay much attention to it. Today, there few systems capture excitement levels and even fewer actually promote awareness exciting moments. In this paper, propose visualization concept for representing individual group-level self-awareness...