D-Ila in V4
PUBLICATIONS
On this page, we summarize and present the results of the D-ILA IN V4 project. Please click on the image of the respective result product to view the displayable data.
Presentation of the data field structure of the D-ILA data model and the possible values of each data field – adult training courses
The table contains the names and characteristics of the data fields selected by our experts. The data fields selected by the experts are part of the D-ILA data model describing adult education. According to the structure defined here, data series edited by experts are produced. The table not only contains the data fields and their characteristics, but also defines precisely the values that each field can take. So, when editing data series, experts can choose from the values specified here. Of course, some of the data fields must be free to complete, such as the cost of training data field.
Presentation of the data field structure of the D-ILA data model and the possible values of each data field – adult learners
The table contains the data fields and their characteristics that – according to our experts – are suitable for describing the characteristics of persons participating in adult education. Similarly to the table of adult trainings, the value sets of each data field are included in this case, which determine what values experts can choose from when editing data series. The last tab of the table (``skipped``) lists data fields that were not included in the data model during expert discussions, but may be relevant for other users.
Data series of the D-ILA data model edited by experts – adult education
According to the original idea, our team of experts was commissioned to define 100 trainings. Finally, the data of 106 trainings were edited. The reason for this is that during the analysis of the training offer, gaps were revealed that were not covered by the 100 trainings, which is why new trainings were included in the table. It is important to note that according to the concept of the project, the trainings in the table – described in the individual data series – are trainings imagined by experts, so it is not possible to enrol in such trainings in practice. On the other hand, the data model and the predefined values of the data fields ensure that real trainings can have the same parameters. Of course, the professional experience of our team of experts also served the purpose of including realistic trainings in the table containing the edited data.
Data series of the D-ILA data model edited by experts – 900 persons participating in adult training
The table contains data of 900 trainees envisioned and edited by experts. Each data series is defined by using the data fields and data field value sets presented in the D-ILA data model. Our experts filled in each data individually to the best of their professional knowledge and experience. This ensures that the data characterizes realistic people. One of the most difficult tasks – needing the most creativity and human resource – of the project was to construct this table.
Data tables for training and testing AI
In accordance with the project plan, the data sets for the description of the 900 trainees produced by our team of experts were divided into two groups. On the one hand, 300 data sets will be used to train the AI, and on the other hand, 600 data series will be used to test the trained AI. In this case, the selection was made using the simplest possible method: each row was assigned an ordinal number using a random number generator, and then, putting the data rows in ascending order based on the random sequence number, we formed a table for training AI from the first 300 data rows. The table accessible via the link already contains the 300 and 600 data series in separate tabs.
Table for editing data of training courses for the development of transversal competences
You can use the table to test how to determine a training course by providing descriptive data. Code tables help fill it in, so you can choose from a drop-down list for each data field. The project experts used a similar data editing interface to construct data from a total of 106 imaginary courses.
WP2 Feasibility Study
This feasibility study is a product of the project Digital Individual Learning Accounts in the Visegrad countries. The project is financed by the European Union.
Interpreting the D-ILA Data Model – presentation
English version. The presentation presents the main features of the Individual Learning Account, the operation of the D-ILA data model and how artificial intelligence-based solutions are applied. Subsequently, the statistical characteristics of the data series generated by experts of trainees and trainings are presented.
Statistical data of 600 adult learners testing trained AI – traditional .pdf report
The opening .pdf is designed for people who do not use Microsoft's PoweBI software. The .pdf file contains a preconfigured, static state of the interactive interface.
Statistical data of 600 adult learners to test trained AI – interactive data visualization (Microsoft PowerBI)
Microsoft PowerBI desktop is required to open the data visualization. This data visualization shows edited data of imaginary individuals that will be used to test the trained AI in WP4. The aim of data visualization is to enable external experts and interested parties to easily and quickly ascertain how realistic the population is based on statistical characteristics.
Statistical data of 300 adult learners for AI training – traditional .pdf statement
The opening .pdf is designed for people who do not use Microsoft's PowerBI software. The .pdf file contains a preconfigured, static state of the interactive interface.
Statistics of 300 adult learners trained in AI – interactive data visualization (Microsoft PowerBI)
Microsoft PowerBI desktop is required to open the data visualization. The project experts constructed the data of a total of 900 imaginary people. Of these, 300 people have been selected to whom experts in WP4 will manually assign training out of the 106 available trainings. The 300 individuals and their assigned trainings will be used to train AI. The data visualization was created to provide an easy and quick overview of the statistical characteristics that allow external experts and interested parties to make sure that the sample created by editing describes a truly realistic population. So AI is trained using data that is realistic.
Statistical data of the 106 adult training data tables – traditional .pdf statement
The opening .pdf is designed for people who do not use Microsoft's PowerBI software. The .pdf file contains a preconfigured, static state of the interactive interface.
D-ILA project – results of quality assurance questionnaire
The questionnaire was completed by 6 people who participated in a workshop to demonstrate the D-ILA data model. Questionnaire respondents do not constitute a representative sample. The graphs in this document were created automatically using Google Forms.
Table for editing data of participants in trainings
You can use the table to test how to determine a person participating in a training by providing descriptive data. Code tables help fill it in, so you can choose from a drop-down list for each data field. The project's experts used a similar data editing interface to construct the data of a total of 900 imaginary individuals.
People participating in adult education – data generated by algorithms
Within the framework of the project, the expert team is primarily responsible for compiling the data series of persons participating in adult education. Nevertheless, we also tried to generate data. The peculiarity of the resulting data series is that the generation was based on a predefined data distribution, i.e. contrary to the data series approach used by experts, in this case the generation took place at the level of data fields. The document contains, on the one hand, the source code of the algorithm used for generation, and on the other hand, the statistical characteristics of the generated sample compared with the statistical characteristics of the data edited by experts.
Adult training courses – data generated by algorithms
In the case of adult training, we experimentally tried the production of data series using an algorithm. In this case, too, the distribution of a data field is the starting point. The document contains, on the one hand, the source code of the algorithm used for generation, and on the other hand, the statistical characteristics of the generated sample compared with the statistical characteristics of the data edited by experts.
D-ILA project – adult training, financing attitude analysis
The questionnaire is available in different languages. The questionnaire includes statements about how to fill in the data series in the ILA data model and how people and training are matched. Statements can be rated on a scale of 1 to 5. As it stands, the questionnaire is not approved by a consortium.
D-ILA project – adult training, funding, attitude test results
The questionnaire was completed by 39 Hungarian persons. 74.4% of the respondents are representatives of adult education organisations. Respondents do not constitute a representative sample. The graphs in this document were created automatically using Google Forms.
D-ILA project – quality assurance questionnaire
English version. The questionnaire is designed for participants in personal consultations and workshops. The questionnaire allows participants to provide feedback on the topics covered in the D-ILA presentation and how to understand them. As it stands, the questionnaire is not approved by a consortium.
Statistics from the 106 adult training course data tables – interactive data visualization (Microsoft PowerBI)
The experts of the project edited the data of 106 trainings. Based on this data, an interactive statistical data visualization was created, which requires the free desktop version of Microsoft PowerBI to be installed on your computer. The goal of data visualization is to make data validation easier. With the help of the presented statistics, external experts who are not familiar with the project can easily and quickly get a comprehensive picture of the training offer consisting of edited trainings. It is important to emphasize that following the objectives of the project, the training offer was compiled by our experts in the most realistic form possible.
