Data Mining Practical Machine Learning Tools and Techniques

(1) Whereas the objectives of the Community, as laid down in the Treaty, as amended by the Treaty on European Union, include creating an ever closer union among the peoples of Europe, fostering closer relations between the States belonging to the Community, ensuring economic and social progress by common action to eliminate the barriers which divide Europe, encouraging the constant improvement of the living conditions of its peoples, preserving and strengthening peace and liberty and promoting democracy on the basis of the fundamental rights recognized in the constitution and laws of the Member States and in the European Convention for the Protection of Human Rights and Fundamental Freedoms;

Some Preliminary Comments on the DIRECTIVE 95/46/EC OF THE EUROPEAN PARLIAMENT AND OF THE COUNCIL of 24 October 1995 on the protection of individuals with regard to the processing of personal data and on the free movement of such data.

The Health Insurance Portability and Accountability Act, also known as HIPAA, was first delivered to congress in 1996 and consisted of just two Titles. It was designed to protect health insurance coverage for workers and their families while between jobs. It establishes standards for electronic health care transactions and addresses the issues of privacy and security when dealing with Protected Health Information (PHI). HIPAA is applicable only in the United States of America.

The Health Insurance Portability and Accountability Act.

Learning Analytics focuses on the collection and analysis of learners' data to improve their learning experience by providing informed guidance and to optimise learning materials. To support the research in this area we have developed a dataset, containing data from courses presented at the Open University (OU). What makes the dataset unique is the fact that it contains demographic data together with aggregated clickstream data of students' interactions in the Virtual Learning Environment (VLE). This enables the analysis of student behaviour, represented by their actions. The dataset contains the information about 22 courses, 32,593 students, their assessment results, and logs of their interactions with the VLE represented by daily summaries of student clicks (10,655,280 entries). The dataset is freely available at https://analyse.kmi.open.ac.uk/open_dataset under a CC-BY 4.0 license.

/pdf/open-university-learning-analytics-dataset-1j54hr1j8j.pdf

Open University Learning Analytics dataset.

Using large, multi-national datasets for high-performance medical imaging AI systems requires innovation in privacy-preserving machine learning so models can train on sensitive data without requiring data transfer. Here we present PriMIA (Privacy-preserving Medical Image Analysis), a free, open-source software framework for differentially private, securely aggregated federated learning and encrypted inference on medical imaging data. We test PriMIA using a real-life case study in which an expert-level deep convolutional neural network classifies paediatric chest X-rays; the resulting model’s classification performance is on par with locally, non-securely trained models. We theoretically and empirically evaluate our framework’s performance and privacy guarantees, and demonstrate that the protections provided prevent the reconstruction of usable data by a gradient-based model inversion attack. Finally, we successfully employ the trained model in an end-to-end encrypted remote inference scenario using secure multi-party computation to prevent the disclosure of the data and the model. Gaining access to medical data to train AI applications can present problems due to patient privacy or proprietary interests. A way forward can be privacy-preserving federated learning schemes. Kaissis, Ziller and colleagues demonstrate here their open source framework for privacy-preserving medical image analysis in a remote inference scenario.

/pdf/end-to-end-privacy-preserving-deep-learning-on-multi-qat6j8x5uq.pdf

End-to-end privacy preserving deep learning on multi-institutional medical imaging

Emerging machine learning technologies are beginning to transform medicine and healthcare and could also improve the diagnosis and treatment of rare diseases. Currently, there are no systematic reviews that investigate, from a general perspective, how machine learning is used in a rare disease context. This scoping review aims to address this gap and explores the use of machine learning in rare diseases, investigating, for example, in which rare diseases machine learning is applied, which types of algorithms and input data are used or which medical applications (e.g., diagnosis, prognosis or treatment) are studied. Using a complex search string including generic search terms and 381 individual disease names, studies from the past 10 years (2010–2019) that applied machine learning in a rare disease context were identified on PubMed. To systematically map the research activity, eligible studies were categorized along different dimensions (e.g., rare disease group, type of algorithm, input data), and the number of studies within these categories was analyzed. Two hundred eleven studies from 32 countries investigating 74 different rare diseases were identified. Diseases with a higher prevalence appeared more often in the studies than diseases with a lower prevalence. Moreover, some rare disease groups were investigated more frequently than to be expected (e.g., rare neurologic diseases and rare systemic or rheumatologic diseases), others less frequently (e.g., rare inborn errors of metabolism and rare skin diseases). Ensemble methods (36.0%), support vector machines (32.2%) and artificial neural networks (31.8%) were the algorithms most commonly applied in the studies. Only a small proportion of studies evaluated their algorithms on an external data set (11.8%) or against a human expert (2.4%). As input data, images (32.2%), demographic data (27.0%) and “omics” data (26.5%) were used most frequently. Most studies used machine learning for diagnosis (40.8%) or prognosis (38.4%) whereas studies aiming to improve treatment were relatively scarce (4.7%). Patient numbers in the studies were small, typically ranging from 20 to 99 (35.5%). Our review provides an overview of the use of machine learning in rare diseases. Mapping the current research activity, it can guide future work and help to facilitate the successful application of machine learning in rare diseases.

/pdf/the-use-of-machine-learning-in-rare-diseases-a-scoping-51yvsfunuw.pdf

The use of machine learning in rare diseases: a scoping review

The sharing of sensitive personal data has become a core element of biomedical research. To protect privacy, a broad spectrum of techniques must be implemented, including data anonymization. In this article, we present ARX, an anonymization tool for structured data which supports a broad spectrum of methods for statistical disclosure control by providing (1) models for analyzing re-identification risks, (2) risk-based anonymization, (3) syntactic privacy criteria, such as k-anonymity, l-diversity, t-closeness and δ-presence, (4) methods for automated and manual evaluation of data utility, and (5) an intuitive coding model using generalization, suppression and microaggregation. ARX is highly scalable and allows for anonymizing datasets with several millions of records on commodity hardware. Moreover, it offers a comprehensive graphical user interface with wizards and visualizations that guide users through different aspects of the anonymization process. ARX is not just a toolbox, but a fully-fledged application, meaning that all implemented methods have been harmonized and integrated with each other. It is well understood that balancing privacy and data utility requires user feedback. To facilitate this interaction, ARX is highly configurable and provides various methods for exploring the solution space.

Putting Statistical Disclosure Control into Practice: The ARX Data Anonymization Tool

Collaboration and data sharing have become core elements of biomedical research. Especially when sensitive data from distributed sources are linked, privacy threats have to be considered. Statistical disclosure control allows the protection of sensitive data by introducing fuzziness. Reduction of data quality, however, needs to be balanced against gains in protection. Therefore, tools are needed which provide a good overview of the anonymization process to those responsible for data sharing. These tools require graphical interfaces and the use of intuitive and replicable methods. In addition, extensive testing, documentation and openness to reviews by the community are important. Existing publicly available software is limited in functionality, and often active support is lacking. We present ARX, an anonymization tool that i) implements a wide variety of privacy methods in a highly efficient manner, ii) provides an intuitive cross-platform graphical interface, iii) offers a programming interface for integration into other software systems, and iv) is well documented and actively supported.

ARX--A Comprehensive Tool for Anonymizing Biomedical Data.

Introduction: This article is part of the Focus Theme of Methods of Information in Medicine on the German Medical Informatics Initiative Future medicine will be predictive, preventive, personalized, participatory and digital Data and knowledge at comprehensive depth and breadth need to be available for research and at the point of care as a basis for targeted diagnosis and therapy Data integration and data sharing will be essential to achieve these goals For this purpose, the consortium Data Integration for Future Medicine (DIFUTURE) will establish Data Integration Centers (DICs) at university medical centers Objectives: The infrastructure envisioned by DIFUTURE will provide researchers with cross-site access to data and support physicians by innovative views on integrated data as well as by decision support components for personalized treatments The aim of our use cases is to show that this accelerates innovation, improves health care processes and results in tangible benefits for our patients To realize our vision, numerous challenges have to be addressed The objective of this article is to describe our concepts and solutions on the technical and the organizational level with a specific focus on data integration and sharing Governance and Policies: Data sharing implies significant security and privacy challenges Therefore, state-of-the-art data protection, modern IT security concepts and patient trust play a central role in our approach We have established governance structures and policies safeguarding data use and sharing by technical and organizational measures providing highest levels of data protection One of our central policies is that adequate methods of data sharing for each use case and project will be selected based on rigorous risk and threat analyses Interdisciplinary groups have been installed in order to manage change Architectural Framework and Methodology: The DIFUTURE Data Integration Centers will implement a three-step approach to integrating, harmonizing and sharing structured, unstructured and omics data as well as images from clinical and research environments First, data is imported and technically harmonized using common data and interface standards (including various IHE profiles, DICOM and HL7 FHIR) Second, data is preprocessed, transformed, harmonized and enriched within a staging and working environment Third, data is imported into common analytics platforms and data models (including i2b2 and tranSMART) and made accessible in a form compliant with the interoperability requirements defined on the national level Secure data access and sharing will be implemented with innovative combinations of privacy-enhancing technologies (safe data, safe settings, safe outputs) and methods of distributed computing Use Cases: From the perspective of health care and medical research, our approach is disease-oriented and use-case driven, ie following the needs of physicians and researchers and aiming at measurable benefits for our patients We will work on early diagnosis, tailored therapies and therapy decision tools with focuses on neurology, oncology and further disease entities Our early uses cases will serve as blueprints for the following ones, verifying that the infrastructure developed by DIFUTURE is able to support a variety of application scenarios Discussion: Own previous work, the use of internationally successful open source systems and a state-of-the-art software architecture are cornerstones of our approach In the conceptual phase of the initiative, we have already prototypically implemented and tested the most important components of our architecture

/pdf/data-integration-for-future-medicine-difuture-o84nmz85qw.pdf

Data Integration for Future Medicine (DIFUTURE).

K-anonymization is an important technique for the de-identification of sensitive datasets. In this paper, we briefly describe an implementation framework which has been carefully engineered to meet the needs of an important class of k-anonymity algorithms. We have implemented and evaluated two major well-known algorithms within this framework and show that it allows for highly efficient implementations. Regarding their runtime behaviour, we were able to closely reproduce the results from previous publications but also found some algorithmic limitations. Furthermore, we propose a new algorithm that achieves very good performance by implementing a novel strategy and exploiting different aspects of our implementation framework. In contrast to the current state-of-the-art, our algorithm offers algorithmic stability, with execution time being independent of the actual representation of the input data. Experiments with different real-world datasets show that our solution clearly outperforms the previous algorithms.

Fabian Prasser

Papers

The use of machine learning in rare diseases: a scoping review

Putting Statistical Disclosure Control into Practice: The ARX Data Anonymization Tool

ARX--A Comprehensive Tool for Anonymizing Biomedical Data.

Data Integration for Future Medicine (DIFUTURE).

Flash: Efficient, Stable and Optimal K-Anonymity