Special Issue "Machine Learning with Applications: Dealing with Interpretability and Imbalanced Datasets"
Deadline for manuscript submissions: 31 October 2021.
Interests: artificial intelligence; data science; machine learning; explainable artificial intelligence; explainable machine learning; human-centric AI; trustworthy internet of things systems
Interests: artificial intelligence; knowledge Engineering; internet of things systems; quality management
Special Issues, Collections and Topics in MDPI journals
Interests: artificial intelligence; machine learning; interpretable machine learning; educational data mining; natural language processing; machine translation
Special Issues, Collections and Topics in MDPI journals
A major disadvantage of using machine learning is that insights about the data are hidden in increasingly complex models. Moreover, the best performing models are often ensembles which cannot be interpreted, even if each single model could be interpreted. Explainable Machine Learning (Explanatory Artificial Intelligence, XAI) summarizes the reasons for black-box behaviour with the aim to gain the trust of users.
This Special Issue of Electronics will provide a forum for discussing exciting research on applying Interpretable Machine Learning (IML) methods on data captured by sensors or generated by interaction of users with systems in a variety of domains. Both original research articles and comprehensive review papers are welcome. We invite also submissions dealing with imbalanced classification problem in which the distribution of examples across the known classes is biased or skewed.
Topics of Interest of this Special Issue include, but are not limited to
- Interpretability (intrinsic or post hoc)
- Global model interpretability
- Local model interpretability
- Feature selection techniques
- Imbalanced classification
- Explainable AI decision support systems
- Real-world applications of Interpretable Machine Learning in areas such as:
- Intelligent transportation systems
- Food safety
- Natural Language Processing
- Smart cities
Prof. Dr. Maja Matetic
Prof. Dr. Xiaoshuan Zhang
Dr. Marija Brkić Bakarić
Manuscript Submission Information
Manuscripts should be submitted online at www.mdpi.com by registering and logging in to this website. Once you are registered, click here to go to the submission form. Manuscripts can be submitted until the deadline. All papers will be peer-reviewed. Accepted papers will be published continuously in the journal (as soon as accepted) and will be listed together on the special issue website. Research articles, review articles as well as short communications are invited. For planned papers, a title and short abstract (about 100 words) can be sent to the Editorial Office for announcement on this website.
Submitted manuscripts should not have been published previously, nor be under consideration for publication elsewhere (except conference proceedings papers). All manuscripts are thoroughly refereed through a single-blind peer-review process. A guide for authors and other relevant information for submission of manuscripts is available on the Instructions for Authors page. Electronics is an international peer-reviewed open access semimonthly journal published by MDPI.
Please visit the Instructions for Authors page before submitting a manuscript. The Article Processing Charge (APC) for publication in this open access journal is 1800 CHF (Swiss Francs). Submitted papers should be well formatted and use good English. Authors may use MDPI's English editing service prior to publication or during author revisions.
- Explainable machine learning
- Interpretable machine learning
- Model interpretability
- Model-agnostic techniques
- Trustworthy Internet of Things (IoT) systems
- Rare event prediction
- Extreme event prediction
- Class imbalance
- Feature selection
The below list represents only planned manuscripts. Some of these manuscripts have not been received by the Editorial Office yet. Papers submitted to MDPI journals are subject to peer-review.
Title: Building optimal machine learning model for predicting peach fruit ripeness using sensor data
Authors: Dejan Ljubobratović, Maja Matetić, Marija Brkić Bakarić, Tomislav Jemrić, Marko Vuković
Affiliation: Authors 1-3: Department of Informatics, University of Rijeka, Centre for Artificial Intelligence and Cybersecurity of the University of Rijeka Authors 4-5: Faculty of Agriculture, University of Zagreb
Abstract: Human experts are commonly used to determine different stages of fruit ripening, whether for picking, transport, or consumption. Building a good model based on data collected by sensors satisfies preconditions for automating the whole process, which can greatly increase the speed of fruit selection, and thus reduce fruit losses due to untimely picking or sorting of fruits. Measurements are made on one hundred peaches in order to collect the sensor data needed to predict ripeness. The sensor data included in this study are mass, volume, electrical impedance, and color. The firmness of the peach, which, according to experts, can represent the ripeness of the fruit, is also measured, as well as acidity (TA) and sugar content (TSS). Since determining peach firmness is a slow and destructive process, its accurate prediction based on sensor data would present a significant improvement of various processes. For example, sensors could be used on a factory line to automatically sort peaches into those ready for storage, transportation, and consumption. Using sensor data to determine peach ripeness in the automatic fruit picking process could significantly speed up the picking process and minimize human error. In this paper, the obtained measurements are used for constructing several machine learning models, such as Support Vector Machines (SVM), Random Forest (RF), and Gradient Boosting Machine (GMB), in order to determine the most accurate peach ripeness prediction model and to assess which variables have the most significant impact in the prediction process. The latter can be very challenging in the so-called blackbox models.