NOESIS lexicon

NOESIS is a two-year research project supported by the European Commission and aims to provide a Decision Support tool and data-driven evidence to enable the deployment of a Big Data in Transport ecosystem in Europe, by addressing the associated technological, institutional/legal, business, and policy challenges.

NOESIS lexicon contains descriptions of the terms that are going to be used throughout the NOESIS project.

Definition: A rule-based machine learning method for discovering ‘interesting’ relations between variables in large datasets. It is intended to identify strong rules in datasets by using measures of ‘interestingness’, such as confidence, collective strength, lift etc.

Short definition: A methodology intending to distinguish data instances by using strong rules and measures of ‘interestingness’.

Definition: The term ‘Big Data’ refers to the techniques of advanced analytics (e.g. machine learning, pattern recognition) or a collection of structured and unstructured, large-volume, highly complex, growing data that may be analysed to reveal underlying patterns, understand complex phenomena, trends, and associations within the data.

Short definition: A vast dataset of high complexity and the techniques of advanced analytics used to handle, process and extract patterns from it.

Definition: The main concept underpinning the NOESIS project. It is based on the idea that useful conclusions can be drawn in respect to the potential value generated from Big Data investments and applications in transport, by applying Big Data techniques on the characteristics and information of such investments and applications relating to the wider transportation and implementation context.

Short definition: An alternative conceptual approach aiming at understanding the key factors of Big Data applications by employing a Big Data approach.

Definition: A specific (ICT) technology making use of Big Data which is implemented or utilized to provide an enhanced service.

Definition: An ‘environment’ that fosters the conditions for Big Data to provide efficient real-time solutions to transportation problems, such that the benefits for all travellers, the society in total and the economy are maximised.

Definition: A collection of more than 100 Big Data Transport use cases, combined in a clearly defined and standardized form. It is a principal section of the Decision Support Framework because it allows access to the use cases and provides support regarding the applicability and requirements of a solution.

Short definition: Similar to a big data/machine learning context, the Big Data in Transport Library constitutes the ‘training examples’ which will be fed into the Learning framework in order to estimate the potential of a specific case.

Definition: The investment corresponds to the resources and the associated cost that have to be allocated and used in order to implement a Big Data application in practice.

Definition: A combination of Big Data technologies and applications which could be implemented in a specific transportation context. Such a product can be comprised of different technologies and ICT applications. Big Data products are characterized by the infrastructure/equipment needed, the data, the analytics methods to be used and the applications which are in place.

Short definition: A well-defined and concise mixture of Big Data requirements, tools, methods and applications to be put into context for a specific transportation problem.

Definition: The broad areas/challenges of interest, in the field of transport and logistics (i.e. planning, operations, management and maintenance) which could be potentially benefited from Big Data applications and related Information and Communication (ICT) technologies.

Short definition: Transport and logistics areas or challenges most likely to be benefited from Big Data applications.

Definition: The proposed business models from NOESIS are designed using the ‘business ecosystem approach’. According to that approach, each model is connected to use cases and allows for assessment of the case's challenges as well as of the value creation for all the stakeholders in the transport big data market. Parts that a traditional business model does not include (e.g. innovative sources of funding) are also taken into account. Such models are envisioned to lead to increased revenues, faster Big Data deployment, informed investment decisions and cost-efficient solutions.

Short definition: Business models that generate feasible investments in managing and optimizing transport systems and networks through Big Data.

Definition: A critical methodological point which will determine the subsequent strategy and methodological approach of the project's learning framework according to the available data and the definition of the ranges of values. The different options will be based on assumptions about the expected results.

Short definition: A milestone of the learning framework which will evaluate the available resources and will point towards the necessary steps to complete the framework.

Definition: A fundamental methodology developed in the framework of the NOESIS project, which aims at evaluating and quantifying the magnitude of the impacts of a Big Data application for a given transport use case, by identifying a set of key performance indicators (KPIs) or evaluation criteria (as referred in the NOESIS terminology).

Short definition: The NOESIS methodology which outputs a quantitative measure of how large or small the impacts of a Big Data application will be in a given transport use case based on specific KPIs.

Definition: A component of the Data Benefit Analysis (see DBA above), which investigates and evaluates the availability of data for a particular case/challenge.

Definition: The use of statistics, machine learning and artificial intelligence methods in order to explore and analyse large datasets with the perspective of finding patterns, learn and generate new information based on the data.

Definition: The procedure of taking decisions based on data or data-driven evidence, rather than mere experience. In NOESIS, this will be achieved by investigating the differences between traditional and Big Data analyses, the development of the Decision Support Tool, the impact assessment methodology as well as the development of realistic business models, a technological and a policy-oriented roadmap.

Short definition: The use of data analytics, tools, models and roadmaps to support the use of Big Data for transportation use cases.

Definition: The outcome of the Learning Big Data Framework. It allows the exploration of methodologies, datasets and implementation procedures (in the form of guidelines) in order to deploy and benefit from the use of Big Data, tailored to the needs specified. It is aimed at estimating the overall potential value creation of Big Data in transport.

Short definition: A set of Machine Learning tools and data processing methods to take advantage of Big Data to improve transport performance.

Definition: A set of KPIs used to assess the benefits or costs inflicted from a Big Data implementation on a particular transport use case.

Short definition: The high-level qualitative or quantitative values used to measure the success or failure of a use case.

Definition: The NOESIS methodology aiming to translate KPI scores of use cases to a generalised well-defined socio-economic value by extending the Cost-Benefit Analysis framework.

Definition: A technology that makes use of information handled by computer or communication networks (e.g. the Internet, smartphone apps, wireless networks, social media). The focus of NOESIS is on the use of such technologies for transportation issues.

Definition: The scoring values of the transport use cases, with regards to the evaluation criteria.

Definition: Also referred as Decision Support Framework. The overall analysis which aims to the development of the decision support tool. The analysis includes the use of machine learning algorithms on all related features identified in WP2 which are represented by indicators, and related to the performance variables of transport infrastructure projects.

Short definition: The overall procedure comprising of i) the definition of features and labels and ii) the identification of the appropriate big data techniques in order to analyse them and identify the underlying data patterns.

Definition: The concept of shifting the use of a transport mode from a personal to a service perspective, thus matching transportation providers and users through ICT. Examples of MaaS include companies such as Uber, DriveNow, Lyft and Via.

Definition: Also referred to as the Responsible code of conduct for Big Data management in transport. A set of guidelines that describe big data techniques and procedures successfully implemented in other fields and the required adjustments for their use in transportation cases. Special attention will be given to ethics and privacy legislation for the successful implementation of Big Data.

Short definition: A set of rules outlining responsible and efficient conduct for Big Data transportation research.

Definition: A plan to implement Big Data for the management and optimization of transport systems and networks. One technological and one policy-oriented roadmap will be developed by NOESIS, with the goal of helping policymakers and managing directions of transport companies to anticipate the right steps for the successful (socio-economic) implementation of Big Data solutions. The development of the roadmaps requires the identification of the main obstacles, a definition of a rational process and a suitable timeframe for the implementation of Big Data as well as a case study-based validation.

Short definition: A plan comprising of the necessary steps to successfully implement Big Data for transport cases by taking into account existing problems and defining a suitable application schedule.

Definition: A (new) dataset consisting of the feature vector of a transport case, on which the Learning Framework is used to predict/evaluate the specific labels of the case, hence assessing the case's impact.

Definition: The training dataset consists of the transport use cases included in a standardized format (i.e. the Big Data in Transport Library) which describes the features of each case and the corresponding labels. The training dataset will form the basis for the development of the Learning Big Data Framework which will be used to identify patterns in the data via learning algorithms.

Short definition: A mapping of f1,f2,…,fn features of a transport use case to a label (vector) Y, describing the scores of the use case according to the evaluation criteria.

Definition: Difficulties, limitations or questions related to a specific transportation problem.

Definition: The input variables of the Decision Support tool. These can be described as the influencing elements in achieving the success of the tool and the project in general. Such variables may be institutional/agency capabilities, governance structures, proper risk allocations etc.

Short definition: The ‘independent’ variables to be used by the decision support tool in order to evaluate Big Data transport use cases.

Definition: Examples of implementing a Big Data solution or service for a specific transportation problem. Each use case is characterised by several features including the transport sector (e.g. passenger of freight), transport mode (e.g. air, rail, road or ship), the application area (e.g. urban, regional, national or international), cost-benefit details etc. The total number of use cases will be collected in the ‘Big Data in Transport Library’.

Short definition: Existing and future cases where Big Data have been applied as tool/solution for a transport-related problem along with their transport, data and geographical characteristics. The main content of the ‘Big Data in Transport Library.

Definition: The evaluation methodologies and criteria developed by the Data Benefit Analysis (DBA) and the Impact Assessment Methodology in WP5.

Definition: The impact of a specific Big Data use case in terms of value for public money, user benefits, life-cycle investment and efficiency. It is evaluated through the aforementioned evaluation criteria (i.e. KPIs).