IU Indianapolis ScholarWorks :: Browsing by Author "Chakraborty, Sunandan"

Browsing by Author "Chakraborty, Sunandan"

Now showing 1 - 9 of 9

Bridging The Gap Between Healthcare Providers and Consumers: Extracting Features from Online Health Forum to Meet Social Needs of Patients using Network Analysis and Embedding
(2020-08) Mokashi, Maitreyi; Chakraborty, Sunandan; Jones, Josette; Zheng, Jiaping
Chronic disease patients have to face many issues during and after their treatment. A lot of these issues are either personal, professional, or social in nature. It may so happen that these issues are overlooked by the respective healthcare providers and become major obstacles in the patient’s day-to-day life and their disease management. We extract data from an online health platform that serves as a ‘safe haven’ to the patients and survivors to discuss help and coping issues. This thesis presents a novel approach that acts as the first step to include the social issues discussed by patients on online health forums which the healthcare providers need to consider in order to create holistic treatment plans. There are numerous online forums where patients share their experiences and post questions about their treatments and their subsequent side effects. We collected data from an “Online Breast Cancer Forum”. On this forum, users (patients) have created threads across many related topics and shared their experiences and questions. We connect the patients (users) with the topic in which they have posted by converting the data into a bipartite network and turn the network nodes into a high-dimensional feature space. From this feature space, we perform community detection on the node embeddings to unearth latent connections between patients and topics. We claim that these latent connections, along with the existing ones, will help to create a new knowledge base that will eventually help the healthcare providers to understand and acknowledge the non-medical related issues to a treatment, and create more adaptive and personalized plans. We performed both qualitative and quantitative analysis on the obtained embeddings to prove the superior quality of our approach and its potential to extract more information when compared to other models.
A Co-Training Model with Label Propagation on a Bipartite Graph to Identify Online Users with Disabilities
(AAAI, 2019) Yu, Xing; Chakraborty, Sunandan; Brady, Erin; Human-Centered Computing, School of Informatics and Computing
Collecting data from representative users with disabilities for accessibility research is time and resource consuming. With the proliferation of social media websites, many online spaces have emerged for people with disabilities. The information accumulated in such places is of great value for data collection and participant recruiting. However, there are also many active non-representative users in such online spaces such as medical practitioners, caretakers, or family members. In this work, we introduce a novel co-training model based on the homophily phenomenon observed among online users with the same disability. The model combines a variational label propagation algorithm and a naive Bayes classifier to identify online users who have the same disability. We evaluated this model on a dataset collected from Reddit and the results show improvements over traditional models.
Investigating the Effect of a Digital Doctor on Persuasion
(2021-10) Dai, Zhengyan; MacDormann, Karl F.; Miller, Andrew; Brady, Erin; Chakraborty, Sunandan
The treatment of chronic diseases requires patient adherence to medical advice. Nonadherence worsens health outcomes and increases healthcare costs. Consultations with a virtual physician could increase adherence, given the shortage of healthcare professionals. However, if the virtual physician is a computer animation, acceptance of its advice may be hampered by the uncanny valley effect, a negative affective reaction to human simulations. Two experiments were conducted to investigate the impact of the virtual physician on patients’ adherence. The first study, a 2 ´ 2 ´ 2 between-groups posttestonly experiment, involved 738 participants playing the role of a patient in a hypothetical virtual consultation with a doctor. The consultation varied in the doctor’s Character, Outcome, and Depiction. Character, Outcome, and Depiction were designed to manipulate the doctor’s level of warmth, competence, and realism. The second study, a 2 ´ 5 between-groups experiment, involved 441 participants assuming a patient’s role in a similar hypothetical virtual consultation with a doctor. The experiment varied the doctor’s Character and Depiction. These independent variables were designed to manipulate the doctor’s level of warmth and eeriness. The first study found that warmth and competence increased adherence intention and consultation enjoyment, but realism did not. On the contrary, the computer-animated doctor increased adherence intention and consultation enjoyment significantly more than the doctor portrayed by a human actor. The enjoyment of the animated consultation caused the doctor to appear warmer and more real, compensating for his realism inconsistency. In the second study, Depiction had a nonsignificant effect on adherence intention, even though the computer animated doctor was perceived as eerier than the real human. The low-warmth, high-eeriness doctor prompted heuristic processing of information, while the high-warmth doctor prompted systematic processing. This pattern runs counter to the literature on persuasion. The doctor’s eeriness, measured in a pretest, had no significant effect on adherence intention via the heuristic-systematic model. Although virtual characters can elicit the uncanny valley effect, they were comparable to a real person in increasing adherence intention, adherence and health behavior. This finding should encourage the development and acceptance of virtual consultation to address the shortage of healthcare professionals.
Political Tweets and Mainstream News Impact in India: A Mixed Methods Investigation into Political Outreach
(ACM, 2018-06) Chakraborty, Sunandan; Chandra, Priyank; Pal, Joyojeet; Romero, Daniel M.; Human-Centered Computing, School of Informatics and Computing
Citizens' perception of politicians and political issues is increasingly influenced by social media. However, little is known about the potential of second order effects of social media in parts of the world where the majority of voting citizens are not online. In this paper, we examine whether a politician can move to communicating through social media as their primary means of outreach, and still present their message to the mainstream population through traditional media. By studying of the use of Twitter by Indian Prime Minister Narendra Modi between 2009 and 2015, the second-most followed elected official in the world, we present evidence of the impact of social media on print news. We use computational text mining techniques to automatically identify print news reports that use Modi's tweets as a source, alongside manual qualitative coding of tweets to analyze the role of tweet themes in print news coverage. We conclude that while Modi's social media messaging does get coverage in the print news, it is his more "newsworthy" tweets, such as references to celebrities, other politicians, or major events such as holidays that have a greater likelihood of coverage.
RAPID: DRL-AI: Investigating A Community-Inclusive AI Chatbot to Support Teachers in Developing Culturally Focused and Universally Designed STEM Activities
(2024-09-14) Price, Jeremy; Chakraborty, Sunandan
Research to uncover and build out the initial feature set for a generative AI chatbot to support teachers in developing more culturally responsive and sustaining STEM lesson plans and activities.
Techniques for Improving the Robustness of Visual Analytics
(2024-08) Koonchanok, Ratanond; Reda, Khairi; Chakraborty, Sunandan; Cafaro, Francesco; McCabe, Sean
Interactive visualization systems, such as Tableau, are integral parts of the data analysis workflow. While such tools were built to help analysts perform exploratory data analysis with minimal effort, analysts have also been using them to make statistical inferences (e.g., predicting future trends) based on patterns revealed by the dataset. However, in addition to revealing true patterns, visualizations can also surface noise and other random fluctuations in data, which could lead to spurious discoveries. The latter poses a threat to the trustworthiness of analyses, especially given the increased reliance on visualizations across various domains. My central thesis is that it is possible to reduce the incidence of false discovery by introducing lightweight user interface elements in visualization tools. In particular, I propose eliciting and incorporating analyst beliefs into visualizations as an approach for guarding against spurious patterns and reducing the risk of analysts “overfitting” the data. To study how analysts would respond to such intervention, I first designed an interactive tool that combined visual belief elicitation with traditional visualization functionalities. In a qualitative study with data analysts, the tool appeared to allow users to operationalize their working knowledge into analyses, nudging them to adopt normative analysis practices (e.g., specifying hypotheses before peeking at data). I then conducted a crowdsourced experiment to investigate if this design could indeed help reduce the incidence of false discovery. Compared to a control condition, participants who used our intervention made significantly more accurate inferences and reported fewer false discoveries. Lastly, I investigated the capability of human intuition by comparing inferences from participants against those generated by statistical machines to understand the advantages and limitations of each. Overall, my thesis paves the way toward the development of a robust visual analytics system that facilitates collaborative decision-making processes, leveraging the complementary abilities of humans and machines.
The Cannabis sativa genetics and therapeutics relationship network: automatically associating cannabis-related genes to therapeutic properties through chemicals from cannabis literature
(BMC, 2023-05-30) Jackson, Trever J.; Chakraborty, Sunandan; BioHealth Informatics, School of Informatics and Computing
Background: Understanding the genome of Cannabis sativa holds significant scientific value due to the multi-faceted therapeutic nature of the plant. Links from cannabis gene to therapeutic property are important to establish gene targets for the optimization of specific therapeutic properties through selective breeding of cannabis strains. Our work establishes a resource for quickly obtaining a complete set of therapeutic properties and genes associated with any known cannabis chemical constituent, as well as relevant literature. Methods: State-of-the-art natural language processing (NLP) was used to automatically extract information from many cannabis-related publications, thus producing an undirected multipartite weighted-edge paragraph co-occurrence relationship network composed of two relationship types, gene-chemical and chemical property. We also developed an interactive application to visualize sub-graphs of manageable size. Results: Two hundred thirty-four cannabis constituent chemicals, 352 therapeutic properties, and 124 genes from the Cannabis sativa genome form a multipartite network graph which transforms 29,817 cannabis-related research documents from PubMed Central into an easy to visualize and explore network format. Conclusion: Use of our network replaces time-consuming and labor intensive manual extraction of information from the large amount of available cannabis literature. This streamlined information retrieval process will enhance the activities of cannabis breeders, cannabis researchers, organic biochemists, pharmaceutical researchers and scientists in many other disciplines.
Using Social Media Websites to Support Scenario-Based Design of Assistive Technology
(2020-01) Yu, Xing; Brady, Erin; Palakal, Mathew; Bolchini, Davide; Chakraborty, Sunandan; Hasan, Mohammad
Having representative users, who have the targeted disability, in accessibility studies is vital to the validity of research findings. Although it is a widely accepted tenet in the HCI community, many barriers and difficulties make it very resource-demanding for accessibility researchers to recruit representative users. As a result, researchers recruit non-representative users, who do not have the targeted disability, instead of representative users in accessibility studies. Although such an approach has been widely justified, evidence showed that findings derived from non-representative users could be biased and even misleading. To address this problem, researchers have come up with different solutions such as building pools of users to recruit from. But still, the data is not widely available and needs a lot of effort and resource to build and maintain. On the other hand, online social media websites have become popular in the last decade. Many online communities have emerged that allow online users to discuss health-related subjects, exchange useful information, or provide emotional support. A large amount of data accumulated in such online communities have gained attention from researchers in the healthcare domain. And many researches have been done based on data from social media websites to better understand health problems to improve the wellbeing of people. Despite the increasing popularity, the value of data from social media websites for accessibility research remains untapped. Hence, my work aims to create methods that could extract valuable information from data collected on social media websites for accessibility practitioners to support their design process. First, I investigate methods that enable researchers to effectively collect representative data from social media websites. More specifically, I look into machine learning approaches that could allow researchers to automatically identify online users who have disabilities (representative users). Second, I investigate methods that could extract useful information from user-generated free-text using techniques drawn from the information extraction domain. Last, I explore how such information should be visualized and presented for designers to support the scenario-based design process in accessibility studies.
Using transfer learning-based causality extraction to mine latent factors for Sjögren’s syndrome from biomedical literature
(Cell Press, 2023-09) VanSchaik, Jack T.; Jain, Palak; Rajapuri, Anushri; Cheriyan, Biju; Thyvalikakath, Thankam P.; Chakraborty, Sunandan; Human-Centered Computing, School of Informatics and Computing
Understanding causality is a longstanding goal across many different domains. Different articles, such as those published in medical journals, disseminate newly discovered knowledge that is often causal. In this paper, we use this intuition to build a model that leverages causal relations to unearth factors related to Sjögren's syndrome from biomedical literature. Sjögren's syndrome is an autoimmune disease affecting up to 3.1 million Americans. Due to the uncommon nature of the illness, symptoms across different specialties coupled with common symptoms of other autoimmune conditions such as rheumatoid arthritis, it is difficult for clinicians to diagnose the disease timely. Due to the lack of a dedicated dataset for causal relationships built from biomedical literature, we propose a transfer learning-based approach, where the relationship extraction model is trained on a wide variety of datasets. We conduct an empirical analysis of numerous neural network architectures and data transfer strategies for causal relation extraction. By conducting experiments with various contextual embedding layers and architectural components, we show that an ELECTRA-based sentence-level relation extraction model generalizes better than other architectures across varying web-based sources and annotation strategies. We use this empirical observation to create a pipeline for identifying causal sentences from literature text, extracting the causal relationships from causal sentences, and building a causal network consisting of latent factors related to Sjögren's syndrome. We show that our approach can retrieve such factors with high precision and recall values. Comparative experiments show that this approach leads to 25% improvement in retrieval F1-score compared to several state-of-the-art biomedical models, including BioBERT and Gram-CNN. We apply this model to a corpus of research articles related to Sjögren's syndrome collected from PubMed to create a causal network for Sjögren's syndrome. The proposed causal network for Sjögren's syndrome will potentially help clinicians with a holistic knowledge base for faster diagnosis.

Browsing by Author "Chakraborty, Sunandan"

Results Per Page

Sort Options