Pub Date : 2022-07-01DOI: 10.1109/IV56949.2022.00015
Maria D'Amaral Ferreira, João Moura Pires, C. Damásio
Ahstract- The occurrence of seasonal natural phenomena depends on the conditions leading to it and not directly on the progression of time, meaning its context varies across time and space. Examples of this include comparing plant growth, insect development or wildfire risk during the same time period at different locations or in different time periods at the same location. However, visualizing and comparing such phenomena usually implies plotting it across the time axis as it's perceived as temporal data. Since it's not directly dependent of time, identifying patters of recurrence using this technique is inefficient. Because of this, we proposed transforming (when needed) the dependent function to a non-decreasing monotone one, in order to preserve the monotonic property of time progression. Then we used the resulting function as a time axis replacement to achieve an equal ground of comparison between the different contexts in which the phenomenon occurs. We applied this technique to real data from seasonal natural phenomena, such as plant and insect growth, to compare its progression in different temporal and spatial contexts. Since the dependent function of the phenomenon was scientifically known, we were able to directly use the technique to infer its seasonality patterns. Furthermore, we applied the technique to real data from the coronavirus worldwide pandemic by hypothesizing its dependent function and analysing if it was able to reduce the existing temporal misalignment between different contexts, like years and countries. The results achieved were positive, although not as remarkable as when the dependent function was known.
{"title":"Visualizing Temporal Data using Time-dependent Non-decreasing Monotone Functions","authors":"Maria D'Amaral Ferreira, João Moura Pires, C. Damásio","doi":"10.1109/IV56949.2022.00015","DOIUrl":"https://doi.org/10.1109/IV56949.2022.00015","url":null,"abstract":"Ahstract- The occurrence of seasonal natural phenomena depends on the conditions leading to it and not directly on the progression of time, meaning its context varies across time and space. Examples of this include comparing plant growth, insect development or wildfire risk during the same time period at different locations or in different time periods at the same location. However, visualizing and comparing such phenomena usually implies plotting it across the time axis as it's perceived as temporal data. Since it's not directly dependent of time, identifying patters of recurrence using this technique is inefficient. Because of this, we proposed transforming (when needed) the dependent function to a non-decreasing monotone one, in order to preserve the monotonic property of time progression. Then we used the resulting function as a time axis replacement to achieve an equal ground of comparison between the different contexts in which the phenomenon occurs. We applied this technique to real data from seasonal natural phenomena, such as plant and insect growth, to compare its progression in different temporal and spatial contexts. Since the dependent function of the phenomenon was scientifically known, we were able to directly use the technique to infer its seasonality patterns. Furthermore, we applied the technique to real data from the coronavirus worldwide pandemic by hypothesizing its dependent function and analysing if it was able to reduce the existing temporal misalignment between different contexts, like years and countries. The results achieved were positive, although not as remarkable as when the dependent function was known.","PeriodicalId":153161,"journal":{"name":"2022 26th International Conference Information Visualisation (IV)","volume":"21 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133339126","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-07-01DOI: 10.1109/IV56949.2022.00056
Joao Vaz, Nuno Datia, Matilde Pato, J. Pires
Road traffic inside cities is responsible for noise and pollution, that causes health problems, fuel consumption and waste of time in jams. Mitigation solutions are usually used to soften the impact of this problem in most cities. In particular, the city of Lisbon has taken measures to reduce pollution by closing areas of the city to the most polluting cars - the zero emission zones. However, the city still lacks visual analytics support for traffic decisions in real-time. In this paper we present a traffic flow indicator that can indicate the road traffic fluidity inside a region of interest for a given time frame, and integrated it into a interactive dashboard supported by a predictive model. With this solution, decision makers can analyse historical data and predict short-term traffic behaviour.
{"title":"Traffic Flow Indicator: Predicting Jams in a City","authors":"Joao Vaz, Nuno Datia, Matilde Pato, J. Pires","doi":"10.1109/IV56949.2022.00056","DOIUrl":"https://doi.org/10.1109/IV56949.2022.00056","url":null,"abstract":"Road traffic inside cities is responsible for noise and pollution, that causes health problems, fuel consumption and waste of time in jams. Mitigation solutions are usually used to soften the impact of this problem in most cities. In particular, the city of Lisbon has taken measures to reduce pollution by closing areas of the city to the most polluting cars - the zero emission zones. However, the city still lacks visual analytics support for traffic decisions in real-time. In this paper we present a traffic flow indicator that can indicate the road traffic fluidity inside a region of interest for a given time frame, and integrated it into a interactive dashboard supported by a predictive model. With this solution, decision makers can analyse historical data and predict short-term traffic behaviour.","PeriodicalId":153161,"journal":{"name":"2022 26th International Conference Information Visualisation (IV)","volume":"13 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133546089","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-07-01DOI: 10.1109/IV56949.2022.00066
A. Dunn, D. Inkpen, Razvan Andonie
The introduction of Transformer neural networks has changed the landscape of Natural Language Processing (NLP) during the recent years. These models are very complex, and therefore hard to debug and explain. In this context, visual explanation became an attractive approach. The visualization of the path that leads to certain outputs of a model is at the core of visual explanation, as this illuminates the features or parts of the model that may need to be changed to achieve the desired results. In particular, one goal of a NLP visual explanation is to highlight the most significant parts of the text that have the greatest impact on the model output. Several visual explanation methods for NLP models were recently proposed. A major challenge is how to compare the performances of such methods since we cannot simply use the usual classification accuracy measures to evaluate the quality of visualizations. We need good metrics and rigorous criteria to measure how useful the extracted knowledge is for explaining the models. In addition, we want to visualize the differences between the knowledge extracted by different models, in order to be able to rank them. In this paper, we investigate how to evaluate explanations/visualizations resulted from machine learning models for text classification. The goal is not to improve the accuracy of a particular NLP classifier, but to assess the quality of the visualizations that explain its decisions. We describe several methods for evaluating the quality of NLP visualizations, including both automated techniques based on quantifiable measures and subjective techniques based on human judgements.
{"title":"Evaluation of Deep Learning Context-Sensitive Visualization Models","authors":"A. Dunn, D. Inkpen, Razvan Andonie","doi":"10.1109/IV56949.2022.00066","DOIUrl":"https://doi.org/10.1109/IV56949.2022.00066","url":null,"abstract":"The introduction of Transformer neural networks has changed the landscape of Natural Language Processing (NLP) during the recent years. These models are very complex, and therefore hard to debug and explain. In this context, visual explanation became an attractive approach. The visualization of the path that leads to certain outputs of a model is at the core of visual explanation, as this illuminates the features or parts of the model that may need to be changed to achieve the desired results. In particular, one goal of a NLP visual explanation is to highlight the most significant parts of the text that have the greatest impact on the model output. Several visual explanation methods for NLP models were recently proposed. A major challenge is how to compare the performances of such methods since we cannot simply use the usual classification accuracy measures to evaluate the quality of visualizations. We need good metrics and rigorous criteria to measure how useful the extracted knowledge is for explaining the models. In addition, we want to visualize the differences between the knowledge extracted by different models, in order to be able to rank them. In this paper, we investigate how to evaluate explanations/visualizations resulted from machine learning models for text classification. The goal is not to improve the accuracy of a particular NLP classifier, but to assess the quality of the visualizations that explain its decisions. We describe several methods for evaluating the quality of NLP visualizations, including both automated techniques based on quantifiable measures and subjective techniques based on human judgements.","PeriodicalId":153161,"journal":{"name":"2022 26th International Conference Information Visualisation (IV)","volume":"52 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128300081","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-07-01DOI: 10.1109/IV56949.2022.00027
Hugh Garner, S. Fernstad
The Scatter Plot Matrix (SPLOM) and the Parallel Coordinates Plot Matrix (PCPM) are frequently used in exploratory data analysis for multivariate data to explore pairwise relationships, clustering and outliers. The SPLOM and PCPM are complex visualization methods with many potential interactions between data, task and visual representation. While numerous studies exist evaluating the SPLOM and Parallel Coordinates Plot (PCP) there is, to the best of our knowledge, no existing study evaluating the PCPM. This pilot study presents an evaluation of the performance of the SPLOM and PCPM for a set of common explorative tasks and identifies key directions for future work. The overall results indicate a minimal performance difference between the visualization methods for most tasks, but with significant variance between users, interactions between data features and response by method, and strong user preferences depending on task. As such, we recommend careful consideration of the background of potential users when choosing a method, and/or the use of complementary or linked views. Further work is required to understand the particular mechanisms impacting users' highly variable performance with the PCPM.
{"title":"Comparative evaluation of the Scatter Plot Matrix and Parallel Coordinates Plot Matrix","authors":"Hugh Garner, S. Fernstad","doi":"10.1109/IV56949.2022.00027","DOIUrl":"https://doi.org/10.1109/IV56949.2022.00027","url":null,"abstract":"The Scatter Plot Matrix (SPLOM) and the Parallel Coordinates Plot Matrix (PCPM) are frequently used in exploratory data analysis for multivariate data to explore pairwise relationships, clustering and outliers. The SPLOM and PCPM are complex visualization methods with many potential interactions between data, task and visual representation. While numerous studies exist evaluating the SPLOM and Parallel Coordinates Plot (PCP) there is, to the best of our knowledge, no existing study evaluating the PCPM. This pilot study presents an evaluation of the performance of the SPLOM and PCPM for a set of common explorative tasks and identifies key directions for future work. The overall results indicate a minimal performance difference between the visualization methods for most tasks, but with significant variance between users, interactions between data features and response by method, and strong user preferences depending on task. As such, we recommend careful consideration of the background of potential users when choosing a method, and/or the use of complementary or linked views. Further work is required to understand the particular mechanisms impacting users' highly variable performance with the PCPM.","PeriodicalId":153161,"journal":{"name":"2022 26th International Conference Information Visualisation (IV)","volume":"13 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122846821","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-07-01DOI: 10.1109/IV56949.2022.00068
N. Lettieri, Delfina Malandrino, Alfonso Guarino, R. Zaccagnino
The digital evolution of economies and markets brings changes that go largely beyond growth and efficiency. In the gig economy, also fueled by algorithmic management solutions, digital labour platforms (DPLs) raise significant issues that include power asymmetries, new forms of workers' abuse and discrimination, algorithms' opacity and over-control. In such a scenario, while normative frameworks evolve novel safeguards, a crucial challenge is that of feeding public (social, institutional) oversight on the dynamics that, at various levels, affect the gig work world. In this paper, we show how the combination of visual analytics and data-driven heuristics can be used to offer new insights on and promote higher levels of transparency and awareness about the fairness of DPLs' activity seen, in the first place, from the workers' perspective. Solutions will be presented as part of GigAdvisor, an experimental cross-platform application developed within an ongoing research that draws on computational social science methods to enable new critical approaches to the digital economy.
{"title":"The Eye of the Rider. Visualization and data-driven heuristics for the critical analysis of gig economy","authors":"N. Lettieri, Delfina Malandrino, Alfonso Guarino, R. Zaccagnino","doi":"10.1109/IV56949.2022.00068","DOIUrl":"https://doi.org/10.1109/IV56949.2022.00068","url":null,"abstract":"The digital evolution of economies and markets brings changes that go largely beyond growth and efficiency. In the gig economy, also fueled by algorithmic management solutions, digital labour platforms (DPLs) raise significant issues that include power asymmetries, new forms of workers' abuse and discrimination, algorithms' opacity and over-control. In such a scenario, while normative frameworks evolve novel safeguards, a crucial challenge is that of feeding public (social, institutional) oversight on the dynamics that, at various levels, affect the gig work world. In this paper, we show how the combination of visual analytics and data-driven heuristics can be used to offer new insights on and promote higher levels of transparency and awareness about the fairness of DPLs' activity seen, in the first place, from the workers' perspective. Solutions will be presented as part of GigAdvisor, an experimental cross-platform application developed within an ongoing research that draws on computational social science methods to enable new critical approaches to the digital economy.","PeriodicalId":153161,"journal":{"name":"2022 26th International Conference Information Visualisation (IV)","volume":"22 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127499990","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-07-01DOI: 10.1109/IV56949.2022.00058
Vít Rusňák, L. Janečková, Filip Drgon, Anna-Marie Dombajova, Veronika Kudelkova
Cybersecurity incident analysis is an exploratory, data-driven process over records and logs from network monitoring tools. The process is rarely linear and frequently breaks down into multiple investigation branches. Analysts document all the steps and lessons learned and suggest mitigations. However, current tools provide only limited support for analytical provenance. As a result, analysts have to record all the details regarding the performed steps and notes in separate documents. Such a procedure increases their cognitive demands and is naturally error-prone. This paper proposes a conceptual design of the analytical tool implementing means of analytical provenance in cybersecurity incident analysis workflows. We identified the user requirements and designed and implemented a proof of concept prototype application Incident Analyzer. Qualitative feedback from four domain experts confirmed that our approach is promising and can significantly improve current cybersecurity and network incident analysis practices.
{"title":"Improving Cybersecurity Incident Analysis Workflow with Analytical Provenance","authors":"Vít Rusňák, L. Janečková, Filip Drgon, Anna-Marie Dombajova, Veronika Kudelkova","doi":"10.1109/IV56949.2022.00058","DOIUrl":"https://doi.org/10.1109/IV56949.2022.00058","url":null,"abstract":"Cybersecurity incident analysis is an exploratory, data-driven process over records and logs from network monitoring tools. The process is rarely linear and frequently breaks down into multiple investigation branches. Analysts document all the steps and lessons learned and suggest mitigations. However, current tools provide only limited support for analytical provenance. As a result, analysts have to record all the details regarding the performed steps and notes in separate documents. Such a procedure increases their cognitive demands and is naturally error-prone. This paper proposes a conceptual design of the analytical tool implementing means of analytical provenance in cybersecurity incident analysis workflows. We identified the user requirements and designed and implemented a proof of concept prototype application Incident Analyzer. Qualitative feedback from four domain experts confirmed that our approach is promising and can significantly improve current cybersecurity and network incident analysis practices.","PeriodicalId":153161,"journal":{"name":"2022 26th International Conference Information Visualisation (IV)","volume":"144 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123476963","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-07-01DOI: 10.1109/IV56949.2022.00072
M. Trutschl, P. Kilgore, Billy A. Tran, Hyun-Woong Nam, Eric Clifford, Adesewa Akande, U. Cvek
Venn diagrams are a useful method of visualizing Boolean data; however, their data aggregation causes fine detail about the data to be lost. In this paper, we present a method of augmenting Venn diagrams, so that they may depict similarity relationships among individual records in the data using the Self-Organizing Map. We applied this method to a synthetic data set and an empirical proteomics data set. We found that we were able to separate data within each region of the Venn diagram based on dimensional values, and that we can highlight the clustering of $p$-values in the empirical set.
{"title":"VennSOM: A SOM-Assisted Visualization of Binary Data","authors":"M. Trutschl, P. Kilgore, Billy A. Tran, Hyun-Woong Nam, Eric Clifford, Adesewa Akande, U. Cvek","doi":"10.1109/IV56949.2022.00072","DOIUrl":"https://doi.org/10.1109/IV56949.2022.00072","url":null,"abstract":"Venn diagrams are a useful method of visualizing Boolean data; however, their data aggregation causes fine detail about the data to be lost. In this paper, we present a method of augmenting Venn diagrams, so that they may depict similarity relationships among individual records in the data using the Self-Organizing Map. We applied this method to a synthetic data set and an empirical proteomics data set. We found that we were able to separate data within each region of the Venn diagram based on dimensional values, and that we can highlight the clustering of $p$-values in the empirical set.","PeriodicalId":153161,"journal":{"name":"2022 26th International Conference Information Visualisation (IV)","volume":"32 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121509250","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-07-01DOI: 10.1109/IV56949.2022.00046
S. Kernbach, Anja Svetina Nabergoj, Anastasia Liakhavets, Andrei Petukh
Today's rapid-changing environment requires individuals, organizations and society-at-large to faster react, act, and proactively shape the future. Design thinking has become a popular innovation method to help proactively design a future to look forward to. However, many design thinking efforts do not go beyond short-term workshops and mini-projects and neglect the longer-term successful implementation of design thinking in organization which is often challenging. Therefore, this paper aims to shed light on the enablers and barriers of successfully implementing design thinking in organizations by providing a first conceptual overview of the enablers and barriers on an organizational and individual level to inform, educate and motivate researchers, practitioners, and educational institutions to discuss and implement design thinking programs in the future more carefully.
{"title":"Design Thinking at a glance - An overview of models along with enablers and barriers of bringing it to the workplace and life","authors":"S. Kernbach, Anja Svetina Nabergoj, Anastasia Liakhavets, Andrei Petukh","doi":"10.1109/IV56949.2022.00046","DOIUrl":"https://doi.org/10.1109/IV56949.2022.00046","url":null,"abstract":"Today's rapid-changing environment requires individuals, organizations and society-at-large to faster react, act, and proactively shape the future. Design thinking has become a popular innovation method to help proactively design a future to look forward to. However, many design thinking efforts do not go beyond short-term workshops and mini-projects and neglect the longer-term successful implementation of design thinking in organization which is often challenging. Therefore, this paper aims to shed light on the enablers and barriers of successfully implementing design thinking in organizations by providing a first conceptual overview of the enablers and barriers on an organizational and individual level to inform, educate and motivate researchers, practitioners, and educational institutions to discuss and implement design thinking programs in the future more carefully.","PeriodicalId":153161,"journal":{"name":"2022 26th International Conference Information Visualisation (IV)","volume":"5 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122281503","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-06-13DOI: 10.1109/IV56949.2022.00060
B. Kovalerchuk, Elijah McCoy
Developing Machine Learning (ML) algorithms for heterogeneous/mixed data is a longstanding problem. Many ML algorithms are not applicable to mixed data, which include numeric and non-numeric data, text, graphs and so on to generate interpretable models. Another longstanding problem is developing algorithms for lossless visualization of multidimensional mixed data. The further progress in ML heavily depends on success interpretable ML algorithms for mixed data and lossless interpretable visualization of multidimensional data. The later allows developing interpretable ML models using visual knowledge discovery by end-users, who can bring valuable domain knowledge which is absent in the training data. The challenges for mixed data include: (1) generating numeric coding schemes for non-numeric attributes for numeric ML algorithms to provide accurate and interpretable ML models, (2) generating methods for lossless visualization of n-D non-numeric data and visual rule discovery in these visualizations. This paper presents a classification of mixed data types, analyzes their importance for ML and present the developed experimental toolkit to deal with mixed data. It combines the Data Types Editor, VisCanvas data visualization and rule discovery system which is available on GitHub.
{"title":"Explainable Mixed Data Representation and Lossless Visualization Toolkit for Knowledge Discovery","authors":"B. Kovalerchuk, Elijah McCoy","doi":"10.1109/IV56949.2022.00060","DOIUrl":"https://doi.org/10.1109/IV56949.2022.00060","url":null,"abstract":"Developing Machine Learning (ML) algorithms for heterogeneous/mixed data is a longstanding problem. Many ML algorithms are not applicable to mixed data, which include numeric and non-numeric data, text, graphs and so on to generate interpretable models. Another longstanding problem is developing algorithms for lossless visualization of multidimensional mixed data. The further progress in ML heavily depends on success interpretable ML algorithms for mixed data and lossless interpretable visualization of multidimensional data. The later allows developing interpretable ML models using visual knowledge discovery by end-users, who can bring valuable domain knowledge which is absent in the training data. The challenges for mixed data include: (1) generating numeric coding schemes for non-numeric attributes for numeric ML algorithms to provide accurate and interpretable ML models, (2) generating methods for lossless visualization of n-D non-numeric data and visual rule discovery in these visualizations. This paper presents a classification of mixed data types, analyzes their importance for ML and present the developed experimental toolkit to deal with mixed data. It combines the Data Types Editor, VisCanvas data visualization and rule discovery system which is available on GitHub.","PeriodicalId":153161,"journal":{"name":"2022 26th International Conference Information Visualisation (IV)","volume":"369 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-06-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122777909","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-05-09DOI: 10.1109/IV56949.2022.00061
Charles Recaido, B. Kovalerchuk
This paper contributes to interpretable machine learning via visual knowledge discovery in general line coordinates (GLC). The concepts of hyperblocks as interpretable dataset units and general line coordinates are combined to create a visual self-service machine learning model. Dynamic Scaffolding Coordinates as lossless multidimensional coordinate systems are proposed, and their applications as visual models is shown. DSC1 and DSC2 can map multiple dataset attributes to a single two-dimensional (X, Y) Cartesian plane using a graph construction algorithm. The hyperblock analysis was used to determine visually appealing dataset attribute orders and to reduce line occlusion. It is shown that hyperblocks can generalize decision tree rules and a series of DSC1 or DSC2 plots can visualize a decision tree. The DSC1 and DSC2 plots were tested on benchmark datasets from the UCI ML repository. They allowed for visual classification of data. Additionally, areas of hyperblock impurity were discovered and used to establish dataset splits that highlight the upper estimate of worst-case model accuracy to guide model selection for high-risk decision-making. Major benefits of DSC1 and DSC2 is their highly interpretable nature. They allow domain experts to control or establish new machine learning models through visual pattern discovery.
{"title":"Interpretable Machine Learning for Self-Service High-Risk Decision-Making","authors":"Charles Recaido, B. Kovalerchuk","doi":"10.1109/IV56949.2022.00061","DOIUrl":"https://doi.org/10.1109/IV56949.2022.00061","url":null,"abstract":"This paper contributes to interpretable machine learning via visual knowledge discovery in general line coordinates (GLC). The concepts of hyperblocks as interpretable dataset units and general line coordinates are combined to create a visual self-service machine learning model. Dynamic Scaffolding Coordinates as lossless multidimensional coordinate systems are proposed, and their applications as visual models is shown. DSC1 and DSC2 can map multiple dataset attributes to a single two-dimensional (X, Y) Cartesian plane using a graph construction algorithm. The hyperblock analysis was used to determine visually appealing dataset attribute orders and to reduce line occlusion. It is shown that hyperblocks can generalize decision tree rules and a series of DSC1 or DSC2 plots can visualize a decision tree. The DSC1 and DSC2 plots were tested on benchmark datasets from the UCI ML repository. They allowed for visual classification of data. Additionally, areas of hyperblock impurity were discovered and used to establish dataset splits that highlight the upper estimate of worst-case model accuracy to guide model selection for high-risk decision-making. Major benefits of DSC1 and DSC2 is their highly interpretable nature. They allow domain experts to control or establish new machine learning models through visual pattern discovery.","PeriodicalId":153161,"journal":{"name":"2022 26th International Conference Information Visualisation (IV)","volume":"101 9","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-05-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"120971591","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：