Browsing by Author "Hoeber, Orland"

Now showing 1 - 20 of 22

Open Access
A study of drag‐and‐drop query refinement and query history visualization for mobile exploratory search
(Wiley, 2024-12-30) Payandeh, Mohammad Hasan; Hoeber, Orland; Boon, Miriam; Storie, Dale; Ramshaw, Veronica
When undertaking complex search scenarios, the underlying information need cannot be satisfied by finding a single optimal resource; instead, searchers need to engage in exploratory search processes to find multiple resources by itera- tively revising and reformulation their queries. This process of query refinement is particularly challenging when using a mobile device, where typing is difficult. Furthermore, in mobile search contexts interruptions can lead to searchers los- ing track of what they were doing. To address these challenges, we designed a public digital library search interface for mobile devices that includes two novel features: drag-and-drop query refinement and query history visualization. To assess the value of this interface compared to a typical baseline, we conducted a controlled laboratory study with 32 participants that included pursuing complex search scenarios, being interrupted in the midst of the search, and resuming the search after the interruption. While participants took more time, they generated longer queries and reported positive subjective opinions about the usability of the exploratory search and task resumption features, along with a greater increase in certainty. These findings show the value of leveraging new touch- based interaction mechanisms within mobile search contexts, and the benefits that visualization can bring to supporting search task resumption
Open Access
A study on interactive visualization for Web information retrieval.
(Faculty of Graduate Studies and Research, University of Regina, 2007) Hoeber, Orland
Open Access
Calculating and Preserving Star Sets and Star Complements of General Matrices
(Faculty of Graduate Studies and Research, University of Regina, 2017-08-18) Bergen, Ryan Paul; Fallat, Shaun; Meagher, Karen; Hoeber, Orland
This thesis presents several results relating to star sets and star complements of graphs. While a method for calculating star sets and star complements involving pro- jection matrices has been known since their introduction, a second method involving determinants is demonstrated and shown to be equivalent to the rst method. Some of the theory for star sets and star complements is expanded to general diagonalizable matrices, regardless of symmetry. The concept of preserving star sets between two general matrices is introduced and shown to be an equivalence relation, and attempts are made to classify what types of matrices can preserve the star sets of a general matrix. Finally, we determine some results for general matrices that occur when star set preservation overlaps with other equivalence relations.
Open Access
Dilex: Enabling cross-session cross-device academic digital library search
(Faculty of Graduate Studies and Research, University of Regina, 2022-06) Gomes, Sebastian Romy; Hoeber, Orland; Mouhoub, Malek; Cushon, Kate
Academic digital library searchers struggle to conduct complex search tasks that are multifaceted in nature and require multiple sessions to finish. When resuming a suspended search task, academic searchers often spend a significant amount of time reacquainting themselves with their past activities and looking for previously issued queries and saved documents. Changing devices while working on these tasks is influenced by their situation and environment. However, current academic digital library search interfaces lack the much-needed behavioural support to continue working on a complex task over a prolonged period and when changing between devices (e.g., starting a search in a mobile device and continuing it in a desktop environment). To assist academic searchers in conducting complex search tasks over multiple sessions and using multiple devices (mobile and desktop/laptop), a novel academic digital library search interface (Dilex) was designed and implemented in this research. The value of Dilex was assessed using a controlled laboratory study, comparing it with the academic digital library search interface of the University of Regina. The results of this study show that searchers were more engaged while working with Dilex, and they interacted more with previous session activities when resuming the task after a one-week gap. It further shows that participants using Dilex spent more time on the search results pages and had an increased degree of interaction with information and personalization features during the resumed tasks. Overall, the results illustrate that academic searchers were able to use the visualizations and interaction features of Dilex to resume a cross-session academic search task when transitioning from a mobile device to a desktop device in the resumed search.
Open Access
The Discovery and Prediction of Genetic Interactions Using Data Science Genetic Interactions
(Faculty of Graduate Studies and Research, University of Regina, 2020-02) Kumar, Ashwani; Zilles, Sandra; Hoeber, Orland; Yao, JingTao; Cameron, Andrew; Vizeacoumar, Franco
The overarching goal of my thesis work was to develop and use data science techniques in order to discover as well as predict genetic interactions (GIs), i.e., functional interactions between gene pairs) so that novel functional associations between genes and their higher-order organizations (protein complexes, pathways and bioprocesses) could be established. Keeping GIs in the center, the work can be divided into three objectives: create useful data on GIs (research article 1; chapter 2), improve methods for predicting the strength of GIs (research article 2; chapter 3), and transfer GI knowledge from one organism to another (research article 3; chapter 4). In research article 1, we generated two GI networks corresponding to two different cellular growth conditions. More than 140,000 gene pairs were analyzed in both conditions, which led to a large amount of data. A comprehensive computational framework was designed to pre-process, benchmark, generate and validate GI networks. The thus produced GI networks were then exhaustively analyzed computationally to obtain new biological insights. These computational analyses helped form an array of biological hypotheses, some of which were then experimentally validated. Since GIs vary in strength, scoring models are used in order to express the strength of a GI. For more than a decade, the multiplicative model has been the scoring model of choice for all Synthetic Genetic Array (SGA) technology-based projects. However, we believed that a better scoring model could be developed by applying machine learning algorithms. We developed multiple scoring models by applying several ii machine learning-based methods and showed that Gaussian Processes (GP) was able to train a model that outperformed all other models, including the multiplicative model. Our new scoring model can help any SGA technology-based study in achieving a better quantification of GIs. Large-scale studies, like research article 1, are hugely informative. However, they are expensive and take years to complete. Moreover, technological limitations prevent us from conducting such studies in higher-order organisms, such as humans. In research article 3, we propose a machine learning-based computational framework to predict GIs in one organism by exploiting GI information from another organism. We predicted over 4,000 previously unknown human GIs by exploiting orthologous GIs in yeast.
Open Access
Drag-and-drop query refinement and query history visualization for mobile exploratory search
(Faculty of Graduate Studies and Research, University of Regina, 2023-12) Payandeh, Mohammad Hasan; Hoeber, Orland; Uddin, Md Sami
Search plays a crucial role in digital public libraries due to the vast amount of information available to users. These users often require support because their search tasks are inherently complex. Complex search tasks, characterized as ill-defined, open-ended, or multi-faceted, occur when the information need cannot be satisfied by simply finding a single optimal resource; searchers are required to revise, refine, and reformulate their queries multiple times. One challenge that users face is how to refine their queries. This challenge is compounded when using a mobile device, where typing is time-consuming and error-prone. Conducting searches in a mobile context adds yet another challenge: the possibility of being interrupted and losing track of the ongoing task. To address these challenges, a novel digital public library search interface, optimized for mobile device use, was designed with two key features: drag-and-drop query refinement and query history visualization. To assess the value of this interface compared to a baseline interface, a controlled laboratory study was conducted with 32 participants. The data analysis results highlight the value of the novel interface in terms of usability measures for search activities (including query refinement), search performance, usability measures during the resumption of search tasks, and task resumption performance. Additionally, the results indicate a significant boost in certainty about the search topic after using the novel interface, unlike the baseline, where no change in certainty was observed. Overall, this user study provides evidence that drag-and-drop interactions can effectively support mobile users in refining their queries. Additionally, it demonstrates the beneficial use of visualization in displaying query history when users need to resume their search tasks.
Open Access
The Effects of Simulated Interruptions on Mobile Search Tasks
(Journal of the Association for Information Science and Technology, 2021-10) Hoeber, Orland; Harvey, Morgan; Dewan Sagar, Shaheed Ahmed; Pointon, Matthew
While it is clear that using a mobile device can interrupt real-world activities such as walking or driving, the effects of interruptions on mobile device use have been under-studied. We are particularly interested in how the ambient distraction of walking while using a mobile device, combined with the occurrence of simulated interruptions of different levels of cognitive complexity, affect web search activities. We have established an experimental design to study how the degree of cognitive complexity of simulated interruptions influence both objective and subjective search task performance. In a controlled laboratory study (n=27), quantitative and qualitative data were collected on mobile search performance, perceptions of the interruptions, and how participants reacted to the interruptions, using a custom mobile eye-tracking app, a questionnaire, and observations. As expected, more cognitively complex interruptions resulted in increased overall task completion times and higher perceived impacts. Interestingly, the effect on the resumption lag or the actual search performance was not significant, showing the resiliency of people to resume their tasks after an interruption. Implications from this study enhance our understanding of how interruptions objectively and subjectively affect search task performance, motivating the need for providing explicit mobile search support to enable recovery from interruptions.
Open Access
Exploring cultural boundries of sport fandom through critical discourse analysis
(Faculty of Graduate Studies and Research, University of Regina, 2019-07) Sveinson, Katherine Larena; Hoeber, Larena; Riemer, Harold; Fink, Janet; Hoeber, Orland; Federick, Evan
This research aimed to critique the notion that sport solely brings people together by exploring discourses related to the creation of cultural boundaries by sport fans through critical discourse studies (CDS) (van Dijk, 2016; Martin, 2002). The three research questions for this dissertation are: How do sport fan-produced texts on social media create, reinforce, or challenge discourses? How do these discourses contribute to the production of cultural boundaries of sport fandom on Twitter? How does drawing cultural boundaries through discourses represent the simultaneously inclusive and exclusive nature of sport fandom on Twitter? A critical discourse analysis (CDA) methodology was employed, using the Toronto Blue Jays 2017 season as a case. Data was collected via Visual Twitter Analytics (Vista) software (Hoeber, Hoeber, El Meseery, Odoh, & Gopi, 2016). This software collects live tweets based on queries. Using the queries #LetsRise, #BlueJays, and @BlueJays, the terms ‘fan’, ‘fans’, and ‘fandom’ were searched for study one. The terms ‘Pillar’, ‘Superman’, and ‘@KPILLAR4’ were searched for study two. Data were read and re-read to determine patterns in the tweets. Next, tweets that represented the different discourses were selected and analyzed using critical discourse analysis. Data analysis involved Fairclough’s (1995a) three-dimensional framework, which involves text analysis, processing analysis, and social analysis, and van Dijk’s (2016) ideological discourse structures, including polarization, pronouns, emphasis on positive self-description and negative other-description, and norms and values. The first study found that using Twitter, fans both reinforced and undermined cultural boundaries based on ideological discourses of sport fandom, which included discourses of loyalty, unity, and consumption. The second study framed reactions to an athlete’s transgression based on the athlete’s brand image associated with Superman. Discourses of Superman, Justice League, and villains demonstrated the ways that cultural boundaries were reinforced and challenged based on opinions of appropriate language in sport. This work demonstrated the simultaneously fluid and rigid nature in which cultural boundaries are drawn. While sport fandom has been primarily explored from a functionalist perspective, these findings suggest that there is also an exclusionary culture, which provides an alternative perspective of sport fandom and fan behaviour.
Open Access
Facet-Time: Using Information Seeking Strategies to Support Exploration of Time Series Data
(Faculty of Graduate Studies and Research, University of Regina, 2019-12) Butt, Haider Ali; Hoeber, Orland; Hamilton, Howard J.; Maguire, Brien; Oriet, Christopher
The notion of time has important value in people’s lives. By analyzing timeoriented data, we can understand underlying trends in the data. Graphically representing the data is useful for analysis because a single visualization can represent much information. Some tools providing visualizations for time-series data are well supported for lookup tasks; however, their support for exploratory search tasks are minimal. In exploratory search tasks, users may not have a specific search goal or they may have ill-defined information needs. They usually learn by iterating through and interpreting search results and perform investigation over them before integrating the obtained information into their knowledge. In this thesis, an analytical tool (Facet-Time) is proposed that supports users in analyzing temporal data with a specific focus on supporting exploratory search. Facet-Time utilizes information visualization and interactive techniques to help users in finding information they are seeking and satisfying their search goals. It allows users with ambiguous search goals to interactively gain information about the search results. Facet-Time provides information scent support to assist users in choosing their path of exploration while they are defining queries. It also provides the visualization of time-series data where users can analyze the data and also make comparisons of subsets of data. User evaluations were conducted to validate the design variations and features of Facet-Time. The outcome of the evaluations gave mostly mixed results, i.e., some experimental measures have significant value for some variations of Facet- Time and some measures do not have much significance. Improvements were observed in three subjective measures (usefulness, ease of use, and satisfaction) when using the sparklines feature versus when using no sparklines feature.
Open Access
Geo-Coordinated Parallel Coordinates (GCPC): A Case Study of Environmental Data Analysis
(Faculty of Graduate Studies and Research, University of Regina, 2016-03) El Meseery, Maha Mohamed Nabil; Hoeber, Orland; Yang, Xue-Dong; Mouhoub, Malek
The large number of environmental problems faced by society in recent years has driven researchers to collect and study massive amounts of environmental data. Such environmental datasets are often high dimensional and heterogeneous in nature, with complex temporal and geospatial relations. The ability to understand and extract meaningful information from such datasets is an essential step for effective decision making. However, reasoning about the data and discovering knowledge in environmental datasets is a challenging problem due to the complexity of the data. The goal of this research is to investigate techniques to support exploration and analysis of environmental data. Such complex data could be characterized as high dimensional heterogeneous geotemporal data. The research focused on the design, implementation, and study of approaches that facilitate the exploration and understanding of such data. A number of visualization approaches have been developed and studied to support the exploration and analysis among environmental datasets, including parallel coordinate plots, geovisualization, investigative scatterplot, and multiple coordinated views. The result of this work was the development of Geo-Coordinated Parallel Coordinates (GCPC), a geovisual analytics approach designed to support the exploration of complex environmental data. Multiple coordinated views are used to represent the high dimensional, heterogeneous, temporal, and geospatial aspects of the data. The approach uses various interactions and analysis features to support exploring and making sense of the data. Field trials were conducted to validate the benefit of the approach in the analysis of environmental data with experts. Environmental analysts used the system to explore within two real datasets in their domain. The results of the evaluation were very positive in general, which provided evidence of the advantages of using the system in exploration among the complex datasets. Domain experts were able to investigate the relations between multiple heterogeneous factors while remaining aware with the geospatial aspect of the data. However, the environmental analysts saw the system as a preliminary exploration tool rather than an analytical approach.
Open Access
Graduate Student Search Strategies within Academic Digital Libraries
(International Journal on Digital Libraries, 2023) Hoeber, Orland; Storie, Dale
Purpose: When searching within an academic digital library, a variety of information seeking strategies may be employed. The purpose of this study is to determine whether graduate students choose appropriate information seeking strategies for the complexity of a given search scenario, and to explore among other factors that could influence their decisions. Methods: We used a survey method in which participants (n=176) were asked to recall their most recent instance of an academic digital library search session that matched two given scenarios (randomly chosen from four alternatives), and for each scenario identify whether they employed search strategies associated with four different information seeking models. Results: Among the search strategies, only lookup search was used in a manner that was consistent with the complexity of the search scenario. Other factors that influenced the choice of strategy were the discipline of study and the type of academic search training received. Patterns of search tool use with respect to the complexity of the search scenarios were also identified. Conclusion: These findings highlight that not only is it important to train graduate students on how to conduct academic digital library searches, more work is needed to train them on matching the information seeking strategies to the complexity of their search tasks and developing interfaces that guide their search process.}
Open Access
Graphical and semantic extensions of variable elimination in Bayesian Networks.
(Faculty of Graduate Studies and Research, University of Regina, 2013-01) Ahmed, Sultan Uddin; Butz, Cortney J.; Maguire, Brien; Hoeber, Orland
Variable Elimination (VE) is a central component of Bayesian network (BN) inference. Two drawbacks of VE are that it has no graphical depiction and that currently its intermediate distributions have no clear semantics. To address these drawbacks, we suggest two approaches, called Graphical VE (GVE) and Semantic VE (SVE). GVE graphically depicts the BN conditional probability tables (CPTs) as Graphical CPTs (GCPTs). The salient characteristic of GVE is that it explains the intricate mathematical equations and notations using graphs without resorting the numerical values and evaluates the inference technique using Merge and Remove graphical operations. GVE may be useful to introduce BN inference to beginners. On the other hand, SVE is an extension of VE that ensures that its intermediate factors have clear semantics, namely, they are defined with respect to the joint probability distribution. SVE exhibits this favorable property by introducing the concept of a p-segment, which ensures well-defined semantics. When the p-segment condition is not satisfied, Bayes theorem can always be applied, iteratively if needed, until it is satisfied. A novel method is suggested for generating p-segments in a good manner. SVE improves clarity in BN inference by avoiding the term potential, where potential is the biggest hindrance to the comprehension of probabilistic inference in BNs.
Open Access
An Image Generation Methodology for Game Engines in Real-Time Using Generative Deep Learning Inference Frameworks
(Faculty of Graduate Studies and Research, University of Regina, 2021-01) Tilson, Adam Richard; Gelowitz, Craig; Yow, Kin-Choong; Hoeber, Orland
Modern video games require image assets for different uses, including representing surface textures of 3D models, like environments, characters or props. These images are commonly generated using photographic and artistic techniques, including hand-drawing and using real-world photographs with additional modifications. This labour-intensive task is typically undertaken by skilled technical artists. However, recently a subset of unsupervised deep machine learning known as Deep Generative Models, have demonstrated the ability to, after sufficient training from an appropriate dataset, create convincing novel counterfeit images mimicking that dataset. This thesis investigates using Deep Generative Models to create image assets for games in real time, directly in the game engine, on common hardware, comparing generation performance across different contexts. The methodology described leverages machine learning as a novel alternative or complementary workflow to traditional image development methods providing some unique advantages to machine learning models. Before these advantages can be exploited, the methodology first needs to be validated. This thesis validates the methodology through demonstrating several procedures for deploying and accessing trained models in a game engine, using inference frameworks running machine learning libraries on the CPU, on low-level matrix math libraries, and on the GPU using compute shaders. The thesis outlines the usage, advantages, limitations and trade-offs to each approach. Six inference frameworks are compared in terms of instantiation and generation time, for generating counterfeit hand-written digits, using various generative models. Next, the thesis investigates a more complicated generation task: generating human faces, and compares performance across engines and platforms. Results show this may be realistically used for real-time generation by the criteria of generating more than ten samples per second in different real-world contexts. Additionally, the generators are confirmed to produce repeatable, consistent results across contexts. Finally, the thesis investigates face generation via disentangled models, observing generated faces may predictably change by traversing the latent space, allowing the model-user to fine tune the generated output. A parameter sweep is performed for the disentanglement hyper-parameter to count how many learned features may be modified independently and predictably, versus reduced reproduction fidelity. The most disentangled model had nine predictably modifiable features.
Open Access
Interactive Image Retrieval, Organization, and Exploration Within Twitter
(Faculty of Graduate Studies and Research, University of Regina, 2020-06) Gaikwad, Manali Vijay; Hoeber, Orland; Yang, Xue Dong; Gelowitz, Craig
Most social media services allow users to share photos. Specifically, with the use of mobile phones, social media users are now able to share photos easily which is leading to the creation of a huge amount of image content. It is difficult to search for specific images within large image collections. Images have textual content associated with them. Users can search for images with textual queries, where the search is performed only on the textual content of the social media post. However, searching within textual information is not sufficient for retrieving images whose visual content is relevant to the query. In this Thesis, ImgSEE (Image SEarch and Exploration) a web-based application for searching and exploring images collected from Twitter is proposed. The system augments the textual contents of the posts with new information extracted from the images using image processing and deep learning methods, and an interactive search interface is built which uses the augmented image data. The images are organized based on visual similarity rather than a grid or list, and additional visual representations of supplemental information are provided to enable interactive filtering and exploration. A user study was conducted with 28 participants to collect evidence on how the approach was used in relation to Vakkari’s three-stage model of information seeking. The study also validated the design choices of ImgSEE in comparison with a traditional grid-based image search interface. The results obtained from the study provide empirical evidence confirming usefulness, ease of use, and user satisfaction with ImgSEE in comparison to a traditional grid-based image search interface.
Open Access
iScholar: A Mobile Research Support System
(Faculty of Graduate Studies and Research, University of Regina, 2013-12) Kim, Sojung; Yao, JingTao; Gerhard, David; Hoeber, Orland; Gelowitz, Craig
The daily lives of researchers have been influenced in many ways by technological developments, especially mobile technology. Nevertheless, little attention has been paid to the formal adaptation of mobile technology into scienti fic research. iScholar, a mobile information retrieval support and recommender system for supporting research activities, is presented. The objective of this thesis is to integrate the latest, particularly mobile, technology when building effective research tools. Ubiquitously available connectivity and new features of mobile devices make it possible to create more flexible and portable research support systems. As a result, advanced research support systems will bring more effi ciency and convenience when conducting scientifi c research. Web-based support systems are studied related to two dimensions - application and technology, and the same applies to mobile-based support systems. As such, the structure of this thesis consists of two parts - scientifi c research as an application domain and mobile technology as a technology dimension. Three research frameworks that are taken from previous studies are reviewed to understand the application domain. One particular framework, the research cycle, is selected for further consideration. Possible mobile support for each step of the chosen framework is identifi ed. The second part of the thesis addresses the technical details of how the system is designed and implemented. Literature review is considered as one of the most important research activities because it influences every aspect of the every aspect of the research process. Therefore, iScholar supports most steps of performing a literature review, such as searching and ltering resource databases, getting paper recommendations to read, and getting conferences/journals recommendations to target. The main functionalities of the proposed system are to retrieve publications as relevant as possible, recommend related papers to read, and recommend conferences to attend or submit papers to. By providing a useful tool, it is hoped that the system saves researcher's time and energy in carrying out research. ii
Open Access
Klink Search: Enabling Exploratory Browsing Activities in Digital Libraries
(Faculty of Graduate Studies and Research, University of Regina, 2021-03) Shukla, Soumya Virendra Kumar; Hoeber, Orland; Fan, Lisa; Waytuck, Brett
Complex search tasks within academic digital libraries go beyond simple lookup. They are open-ended, multi-faceted, with ill-defined information needs, requiring cognitive processing of the encountered information and learning. Academic digital library searchers often use exploratory search strategies for complex search tasks. However, the present academic digital library search interfaces provide limited support for exploratory browsing activities, which is a crucial step that supports learning during the exploratory search process. The present research addresses this issue by designing and implementing a novel search interface called KLink Search, to support exploratory browsing activities via two new features: visually linked keywords and interactive workspace. KLink Search leverages information visualization principles to enable interactive information retrieval in academic digital libraries. A controlled laboratory study was conducted with 32 participants to gather evidence for the value of KLink Search compared to a Baseline interface. The data analysis results show that the participants reported higher degrees of perceived ease-of-use, usefulness, satisfaction, and knowledge gain after using KLink Search for complex search tasks. Further, they spend more time searching for information and saving a highly relevant set of documents at the end of the tasks when using KLink Search. These results depict the value of providing lightweight visual additions to the existing academic digital library search interfaces to support exploratory browsing activities effectively.
Open Access
Real-Time Sentiment-Based Anomaly Detection in Twitter Data Streams
(Faculty of Graduate Studies and Research, University of Regina, 2016-03) Patel, Khantil Ragnesh; Hamilton, Howard J.; Hoeber, Orland; Hilderman, Robert
Twitter has over 316 million active users and the engagement of these Twitter users results in the rapid production of data, notably in the context of popular topics (such as news stories, politics, and sports). This data is available in the form of data streams, which has led many researchers to develop analysis techniques especially for Twitter data streams. Although anomaly detection in time series is a well established research area, its application to detect sentiment-based anomalies in large volumes of streaming data began recently. A sentiment-based anomaly is de ned as a sudden increase in the time series of tweets individually associated with a positive, neutral, or negative sentiment. The goal of this research is to develop and evaluate a technique to automatically detect sentiment-based anomalies, while avoiding the repeated detection of anomalies of similar types. Detecting anomalies in data streams is challenging due the requirement that anomalies be detected in real-time. We propose an approach for real-time sentiment-based anomaly detection (RSAD) in Twitter data streams. Sentiment classi cation is used to split the input data stream into three independent streams (positive, neutral, and negative), which are then analyzed separately for anomalous spikes in the number of tweets. Rare anomalies and the rst occurrence of repeated anomalies are distinguished from the repeated occurrence of similar anomalies. Six approaches for anomaly detection in data streams, including two baseline approaches, are described. These approaches were tested on two user-generated datasets. The rst dataset concerned an international sports event and was collected from Twitter and the second concerned a political party and was collected from multiple social media platforms. Results from these evaluations show that a probabilistic exponentially weighted moving average (PEWMA), coupled with a sliding window that uses a median absolute deviation (MAD) calculation, is effective at identifying sentiment-based anomalies. The PEWMA-MAD approach is consistently among the top two methods for all cases tested. The simple linear regression approach is slightly better in the case of the second dataset. Overall, the results suggest that the PEWMA-MAD approach may be robust su ciently to be applied to a wide variety of datasets from di erent social media platforms. ii
Open Access
Search Timelines: Contextualized search history in support of exploratory search
(Faculty of Graduate Studies and Research, University of Regina, 2024-06) Islam, MD Nazmul; Hoeber, Orland; Uddin, Md Sami
The timespan over which exploratory searching can occur, as well as the scope and volume of the search activities undertaken, can make it difficult for searchers to remember key details about their complex search tasks. These difficulties are present both during a single search session and when resuming a search task that spans multiple sessions. Existing public digital library search interfaces do not provide adequate support to allow patrons to pursue a complex search task for a prolonged period across multiple sessions. In this thesis, I present a novel search interface designed to support cross-session exploratory search in a public digital library context through the introduction of Search Timelines. The interface provides context for past and current search activities via a dynamically updating timeline of the search (queries and saved documents). This timeline is presented alongside the search results in a typical search engine results page design. A more detailed timeline is provided in a workspace, where searchers can review the history of their search activities and their saved resources. In order to assess the value of Search Timelines, a controlled laboratory study was conducted to compare it with a baseline interface modelled after the Regina Public Library search interface. The results of this study show that searchers were more engaged while working with Search Timelines both during the initial search session and when resuming the search after a gap of 7-8 days. The results also show the value of the novel interface in terms of usability measures, perceived knowledge gain, and search performance for both initial and resumed sessions. Overall, the results illustrate that Search Timelines can effectively support public digital library searchers while performing exploratory search that span multiple sessions.
Open Access
Sparse Coding Tone-Like Structures in Sound Using Local Image Features
(Faculty of Graduate Studies and Research, University of Regina, 2015-08) Ubbens, Jordan Robert; Gerhard, David; Yang, Xue-Dong; Hoeber, Orland; Gelowitz, Craig
A trend in machine learning has emphasized the use of features which are learned algorithmically, in contrast to the hand-engineered features traditionally used in classification tasks. Classical sparse coding is a robust feature learning paradigm which represents inputs as a sparse vector of coefficients applied to a dictionary of basis functions. While sparse coding has yielded state-of-the-art results in many application domains, computational challenges often make it impractical to use. This thesis examines the application of a local image feature based sparse coding algorithm (ScSPM) to the problem domain of audio classification. The convex optimization problems involved in dictionary learning are discussed, and existing methods are reviewed. With the goal of mitigating some of the computational expense involved in sparse coding local image features, alternative image-based representations of audio are proposed which isolate the tone-like structures present in the signal. The proposed alternative representations are evaluated in a multi-class audio classification task with respect to training time as well as classification accuracy. i
Open Access
A Study of Visually Linked Keywords to Support Exploratory Browsing in Academic Search
(Journal of the Association for Information Science and Technology, 2022) Hoeber, Orland; Shukla, Soumya
While the search interfaces used by common academic digital libraries pro- vide easy access to a wealth of peer-reviewed literature, their interfaces pro- vide little support for exploratory browsing. When faced with a complex search task (such as one that requires knowledge discovery), exploratory browsing is an important first step in an exploratory search process. To more effectively support exploratory browsing, we have designed and implemented a novel academic digital library search interface (KLink Search) with two new features: visually linked keywords and an interactive workspace. To study the potential value of these features, we have conducted a controlled laboratory study with 32 participants, comparing KLink Search to a baseline digital library search interface modelled after that used by IEEE Xplore. Based on subjective opinions, objective performance, and behavioural data, we show the value of adding lightweight visual and interactive features to academic digital library search interfaces to support exploratory browsing.