scholarly journals Investigation of Combining Logitboost(M5P) under Active Learning Classification Tasks

Informatics ◽  
2020 ◽  
Vol 7 (4) ◽  
pp. 50
Author(s):  
Vangjel Kazllarof ◽  
Stamatis Karlos ◽  
Sotiris Kotsiantis
Keyword(s):  
Active Learning ◽  
Real Life ◽  
Regression Tree ◽  
Predictive Ability ◽  
Powerful Learning ◽  
Robust Decisions ◽  
Classification Tasks ◽  
Weakly Supervised ◽  
Boosting Technique ◽  
Base Learner

Active learning is the category of partially supervised algorithms that is differentiated by its strategy to combine both the predictive ability of a base learner and the human knowledge so as to exploit adequately the existence of unlabeled data. Its ambition is to compose powerful learning algorithms which otherwise would be based only on insufficient labelled samples. Since the latter kind of information could raise important monetization costs and time obstacles, the human contribution should be seriously restricted compared with the former. For this reason, we investigate the use of the Logitboost wrapper classifier, a popular variant of ensemble algorithms which adopts the technique of boosting along with a regression base learner based on Model trees into 3 different active learning query strategies. We study its efficiency against 10 separate learners under a well-described active learning framework over 91 datasets which have been split to binary and multi-class problems. We also included one typical Logitboost variant with a separate internal regressor for discriminating the benefits of adopting a more accurate regression tree than one-node trees, while we examined the efficacy of one hyperparameter of the proposed algorithm. Since the application of the boosting technique may provide overall less biased predictions, we assume that the proposed algorithm, named as Logitboost(M5P), could provide both accurate and robust decisions under active learning scenarios that would be beneficial on real-life weakly supervised classification tasks. Its smoother weighting stage over the misclassified cases during training as well as the accurate behavior of M5P are the main factors that lead towards this performance. Proper statistical comparisons over the metric of classification accuracy verify our assumptions, while adoption of M5P instead of weak decision trees was proven to be more competitive for the majority of the examined problems. We present our results through appropriate summarization approaches and explanatory visualizations, commenting our results per case.

A deep metric learning approach for weakly supervised loan default prediction1

2021 ◽  
pp. 1-13
Author(s):  
Kai Zhuang ◽  
Sen Wu ◽  
Xiaonan Gao
Keyword(s):  
Default Risk ◽  
Metric Learning ◽  
Real Life ◽  
Risk Scores ◽  
Loan Default ◽  
Default Prediction ◽  
Heterogeneous Features ◽  
Deep Metric Learning ◽  
Weakly Supervised ◽  
Decision Boundaries

To deal with the systematic risk of financial institutions and the rapid increasing of loan applications, it is becoming extremely important to automatically predict the default probability of a loan. However, this task is non-trivial due to the insufficient default samples, hard decision boundaries and numerous heterogeneous features. To the best of our knowledge, existing related researches fail in handling these three difficulties simultaneously. In this paper, we propose a weakly supervised loan default prediction model WEAKLOAN that systematically solves all these challenges based on deep metric learning. WEAKLOAN is composed of three key modules which are used for encoding loan features, learning evaluation metrics and calculating default risk scores. By doing so, WEAKLOAN can not only extract the features of a loan itself, but also model the hidden relationships in loan pairs. Extensive experiments on real-life datasets show that WEAKLOAN significantly outperforms all compared baselines even though the default loans for training are limited.

scholarly journals Evolutionary Multitasking-Based Multiobjective Optimization Algorithm for Channel Selection in Hybrid Brain Computer Interfacing Systems

2021 ◽  
Vol 15 ◽  
Author(s):  
Tianyu Liu ◽  
Zhixiong Xu ◽  
Lei Cao ◽  
Guowei Tan
Keyword(s):  
Multiobjective Optimization ◽  
Real Life ◽  
Channel Selection ◽  
Research Field ◽  
Neural Engineering ◽  
Computer Interfaces ◽  
Classification Tasks ◽  
Optimization Methodology ◽  
Brain Computer Interfacing ◽  
Selection Of

Hybrid-modality brain-computer Interfaces (BCIs), which combine motor imagery (MI) bio-signals and steady-state visual evoked potentials (SSVEPs), has attracted wide attention in the research field of neural engineering. The number of channels should be as small as possible for real-life applications. However, most of recent works about channel selection only focus on either the performance of classification task or the effectiveness of device control. Few works conduct channel selection for MI and SSVEP classification tasks simultaneously. In this paper, a multitasking-based multiobjective evolutionary algorithm (EMMOA) was proposed to select appropriate channels for these two classification tasks at the same time. Moreover, a two-stage framework was introduced to balance the number of selected channels and the classification accuracy in the proposed algorithm. The experimental results verified the feasibility of multiobjective optimization methodology for channel selection of hybrid BCI tasks.

Teaching sustainability-oriented capabilities using active learning approach

2021 ◽  
Vol ahead-of-print (ahead-of-print) ◽  
Author(s):  
Priscila Borin Claro ◽  
Nathalia Ramajo Esteves
Keyword(s):  
Active Learning ◽  
Business School ◽  
Business Schools ◽  
School Context ◽  
Real Life ◽  
Learning Approach ◽  
Case Study Methodology ◽  
Content Type ◽  
Positive Effects ◽  
The Impact

Purpose This paper aims to discuss how educators can teach sustainability-oriented capabilities (SOCs) using an active learning approach. Design/methodology/approach Using a case study methodology centered on a Brazilian business school, this research combines qualitative analysis of content, such as teacher notes and student work, with quantitative analysis of student grades. The authors used variance analysis and Bonferroni tests to establish whether the means of three test groups were significantly different. The authors also tested for normality, using the Skewness Kurtosis test, and for homoscedasticity, using Levene. Findings The authors’ findings suggest that the active learning (AL) method may be useful in developing SOCs related to the capabilities of “to know,” “to do,” “to interact” and “to be” because it improved student engagement in the program. In addition, this improved engagement was shown to have a positive influence on grades. Research limitations/implications Using convenience sampling, the authors studied a limited number of the mandatory management courses offered by Insper. There is a need to check for nonlinear positive effects over a more extended period of time and considering more courses. Practical implications This paper offers a practical and replicable technique for teaching SOCs in a business school context using AL. Originality/value The existing literature on education and sustainability discusses the role of business schools in the development of SOCs, especially with respect to curricular changes that integrate content and frameworks related to the conceptualization of sustainable development for business (Cebrián and Junyent, 2015; Cortese, 2003; Fairfield, 2018; Aleixo et al., 2020; Leal Filho, 2020; Arruda Filho et al., 2019). However, some studies suggest that the learning process at many business schools fails to explore the complexity of real life by not using a teaching approach that favors the development of SOCs (Leal Filho et al., 2015). Thus, prior studies have pointed to the need for further research on the impact of the active learning approach in teaching about sustainability (Leal Filho et al., 2015; Fisher and Bonn, 2011; Hesselbarth and Schaltegger, 2014). The aim of this research is to contribute to this discussion.

scholarly journals Motivation, Engagement and Learning through Digital Games

2011 ◽  
Vol 2 (2) ◽  
pp. 1-16 ◽  
Author(s):  
Ioanna Iacovides ◽  
James Aczel ◽  
Eileen Scanlon ◽  
Josie Taylor ◽  
Will Woods
Keyword(s):  
Active Learning ◽  
Informal Learning ◽  
Learning Environments ◽  
Empirical Studies ◽  
Digital Games ◽  
Powerful Learning ◽  
Educational Environments ◽  
Games For Learning ◽  
Affinity Groups ◽  
Powerful Learning Environments

Digital games can be powerful learning environments because they encourage active learning and participation within “affinity groups” (Gee, 2004). However, the use of games in formal educational environments is not always successful (O’Neil et al., 2005). There is a need to update existing theories of motivation and engagement in order to take recent game-related developments into account. Understanding the links between why people play games, what keeps them engaged in this process, and what they learn as a result could have a significant impact on how people value and use games for learning. This paper examines key research that relates to motivation, engagement, and informal learning through digital games, in order to highlight the need for empirical studies which examine the activities that occur in and around everyday gaming practice.

The Study of Knowledge-Based Active Learning Grid Operation Experience

2013 ◽  
Vol 860-863 ◽  
pp. 2456-2462
Author(s):  
Qiang Qing Zhou ◽  
Jing Liu ◽  
Qing Nian Zou ◽  
Guo Lin Huang ◽  
Ping Han ◽  
...  
Keyword(s):  
Active Learning ◽  
Knowledge Base ◽  
System Structure ◽  
Learning Ability ◽  
Learning Capability ◽  
Knowledge Based ◽  
Object Oriented Technology ◽  
Powerful Learning ◽  
Grid Operation

By using of the object-oriented technology and the knowledge representation of the production rule, this paper classifies the operation experience of grid according to the nature and builds a power grid operation experience knowledge base with active learning capability. Through application of Bayesian classifier model based on weight, it classifies the statistical data and identifies the semantic, to realize the exchange between the knowledge base and the users feedback. Using the powerful learning ability of knowledge base, it can make the operation experience knowledge base optimize its knowledge system structure while exchanging with users feedback, so that it can go on refining the operation experience base of the grid. This method can provide technical support and improve the quality of the stuff, as well as strengthen the security and stability of the grid.

scholarly journals LET THE STUDENTS DESIGN THEIR OWN EXPERIMENTS!

2019 ◽  
Author(s):  
Karen A. Lawrence
Keyword(s):  
Active Learning ◽  
Design Of Experiments ◽  
Learning Outcomes ◽  
Real Life ◽  
Engineering Practice ◽  
Directed Design ◽  
American Society ◽  
Student Reports ◽  
Industry Professionals ◽  
Student Confidence

The objective of this paper is to illustrate the use and benefits of a student-directed Design of Experiments (DOE) project as an active learning instrument within a second course in statistics for students enrolled in one of three programs in the Bachelor of Technology at the W. Booth School of Engineering Practice and Technology, McMaster University. Pedagogy will be considered, learning outcomes presented, level and depth of topic areas will be explored and evidence of benefit to the students will be shared. Examples of student objective statements will be given to show the level of interest in conducting a self-chosen experiment. Concluding comments from student reports will be highlighted to demonstrate how the project serves as a useful vehicle for discussing practicalities that arise in real life investigations. Lastly, details about the yearly American Society for Quality (ASQ) Student Quality Showcase event will be shared to demonstrate how interaction with industry professionals enhances student confidence and develops key attributes desired in engineering professionals.

scholarly journals Promoting an active form of learning out-of-class via answering online “study questions” leads to higher than expected exam scores in General Biology

PeerJ ◽  
2015 ◽  
Vol 3 ◽  
pp. e1322 ◽  
Author(s):  
Susan I. Gibson
Keyword(s):  
Active Learning ◽  
Student Learning ◽  
Learning Outcomes ◽  
Active Form ◽  
Real Life ◽  
Student Learning Outcomes ◽  
General Biology ◽  
Online Study ◽  
Covering Material ◽  
Exam Scores

A rising need for workers in science, technology, engineering and mathematics (STEM) fields has fueled interest in improving teaching within STEM disciplines. Numerous studies have demonstrated the benefits of active learning approaches on student learning outcomes. However, many of these studies have been conducted in experimental, rather than real-life class, settings. In addition, most of these studies have focused on in-class active learning exercises. This study tested the effects of answering questions outside of class on exam performance for General Biology students at the University of Minnesota. An online database of 1,020 multiple-choice questions covering material from the first half of the course was generated. Students in seven course sections (with an average of ∼265 students per section) were given unlimited access to the online study questions. These students made extensive use of the online questions, with students answering an average of 1,323 questions covering material from the half of the semester for which the questions were available. After students answered a set of questions, they were shown the correct answers for those questions. More specific feedback describing how to arrive at the correct answer was provided for the 73% of the questions for which the correct answers were not deemed to be self-explanatory. The extent to which access to the online study questions improved student learning outcomes was assessed by comparing the performance on exam questions of students in the seven course sections with access to the online study questions with the performance of students in course sections without access to the online study questions. Student performance was analyzed for a total of 89 different exams questions that were not included in the study questions, but that covered the same material covered by the study questions. Each of these 89 questions was used on one to five exams given to students in course sections that had access to the online study questions and on three to 77 exams given to students in sections that lacked such access. Data from over 1,800 students in sections with access to the online study questions show that those students scored a statistically significant average of 6.6% points higher on the exam questions analyzed than students in sections without access to the study questions. This difference was greater than the average amount necessary to raise students’ exam grades by one grade (e.g., from a “B-” to a “B”). In addition, there was a higher correlation between number of questions answered and success on exam questions on material related to the study questions than between number of questions answered and success on exam questions on material unrelated to the study questions. The online study question system required substantial effort to set up, but required minimal effort to maintain and was effective in significantly raising average exam scores for even very large course sections.

scholarly journals A Weakly Supervised Method for Mud Detection in Ores Based on Deep Active Learning

2020 ◽  
Vol 2020 ◽  
pp. 1-10
Author(s):  
Zhijian Huang ◽  
Fangmin Li ◽  
Xidao Luan ◽  
Zuowei Cai
Keyword(s):  
Deep Learning ◽  
Active Learning ◽  
Learning Model ◽  
Experimental Results ◽  
Detection Accuracy ◽  
Challenging Problem ◽  
Real Scene ◽  
Weakly Supervised ◽  
Deep Learning Model

Automatically detecting mud in bauxite ores is important and valuable, with which we can improve productivity and reduce pollution. However, distinguishing mud and ores in a real scene is challenging for their similarity in shape, color, and texture. Moreover, training a deep learning model needs a large amount of exactly labeled samples, which is expensive and time consuming. Aiming at the challenging problem, this paper proposed a novel weakly supervised method based on deep active learning (AL), named YOLO-AL. The method uses the YOLO-v3 model as the basic detector, which is initialized with the pretrained weights on the MS COCO dataset. Then, an AL framework-embedded YOLO-v3 model is constructed. In the AL process, it iteratively fine-tunes the last few layers of the YOLO-v3 model with the most valuable samples, which is selected by a Less Confident (LC) strategy. Experimental results show that the proposed method can effectively detect mud in ores. More importantly, the proposed method can obviously reduce the labeled samples without decreasing the detection accuracy.

scholarly journals Impactful engineering education through sustainable energy collaborations with public and private entities

2019 ◽  
Vol 20 (2) ◽  
pp. 393-407
Author(s):  
Lluís Pacheco ◽  
Luo Ningsu ◽  
Toni Pujol ◽  
Jose Ramon Gonzalez ◽  
Inès Ferrer
Keyword(s):  
Renewable Energy ◽  
Active Learning ◽  
Undergraduate Students ◽  
Sustainable Energy ◽  
Renewable Energy Sources ◽  
Real Life ◽  
Energy Sources ◽  
Low Carbon ◽  
Public And Private ◽  
Content Type

Purpose This paper aims to report on a case study concerning the development of sustainable energy partnerships involving engineering faculty and undergraduate students at the University of Girona, Catalonia, Spain. Design/methodology/approach Faculty were motivated to seek partnerships with public and private entities in the local area for the purposes of realising mutually beneficial outcomes. The educational programmes of future engineers, when sustainability is considered, are analysed. Education for sustainable development has to include multidisciplinary active learning as a desirable competence. Active learning can be obtained when problems are based on real life because they are most motivating for students. Constructive alignment component is obtained because learning objectives are linked with learning activities related to the needs of public and private entities. Findings Through the provision of technical expertise, the adoption and success of renewable energy projects was facilitated on the one hand, while final year undergraduate students benefited in terms of hands-on experience in helping to bring these projects to life, drawing on the knowledge and skills they had acquired throughout their degree programmes. These works are addressed to students by faculty members with the aim of developing and promoting renewable energies. Outcomes from partnerships surpassed expectations; not only were different benefits realised as were initially hoped for, but this success led to partnerships being sustained over time. Originality/value Fossil fuel-based energy systems are associated with a myriad of negative environmental and social externalities. It is difficult to overstate the importance of transitioning towards alternative low carbon energy sources for climate change mitigation which are less centralised compared to the status-quo for energy security and energy independence. By actively facilitating the development of decentralised renewable energy sources in Catalonia, the projects reported herein are of significant value in social environmental and educational terms.

Sign in / Sign up

Export Citation Format

Share Document