Tag: text complexity analysis

Year 2 Periodic Report of the Project

Written by iread4skills on March 17, 2025. Posted in News.

iRead4Skills has just released its Year 2 Periodic Report, highlighting significant progress in the development of intelligent tools to improve literacy in adults with low reading skills.

Aiming to make reading more accessible and effective for Adult Learning (AL) and Vocational Education and Training (VET) learners, the initiative has been combining artificial intelligence (AI) and natural language processing (NLP) to classify texts according to their complexity and create innovative reading support solutions.

Key Achievements in the Second Year of the Project

📌 Creation of a specialized linguistic corpus in Portuguese, Spanish, and French, classifying texts by complexity levels for further analysis.
📌 Development of the Intelligent Complexity Analyzer (ICA) API, an innovative system that allows evaluating text readability, identifying complex linguistic patterns, and providing recommendations adapted to users’ literacy levels.
📌 Adoption of machine learning models capable of predicting and automatically classifying the reading difficulty of different types of content.
📌 Co-creation workshops with trainers and learners to validate and improve tools, ensuring they meet the real needs of the target audience.
📌 Strengthened collaboration with policymakers and training institutions, ensuring the practical implementation of developed solutions and their integration into educational programs.
📌 Promotion of Open Science, ensuring that collected data and developed methodologies are accessible to researchers and educators.

The Year 2 Periodic Report is now publicly available and can be accessed via the link: https://zenodo.org/records/15038440

iRead4Skills Presented to High School Students in Galicia

Written by iread4skills on March 10, 2025. Posted in News.

On February 20, 2025, the team from the University of Santiago de Compostela (USC), represented by Marcus Garcia Gonzales and Sandra Rodriguez Rey, presented the iRead4Skills project to 12th-grade high school students in Galicia.

During the session, students were introduced to the tools developed as part of the project, aimed at improving reading and writing skills through artificial intelligence. The presentation included a live demonstration of the Text Classifier by Level of Complexity and the Writing Analyzer/Assistant, allowing participants to explore their potential use in education.

This initiative provided an excellent opportunity to engage students in discussions on technology and education, sparking interest in how digital tools can support learning.

The iRead4Skills project remains committed to educational innovation, integrating feedback from teachers and students to refine its applications.

New iRead4Skills Deliverable: Annotated Corpora by Level of Complexity for FR, PT, and SP

Written by iread4skills on February 5, 2025. Posted in News.

We are pleased to announce the release of Dataset 2: Annotated Corpora by Level of Complexity for French (FR), Portuguese (PT), and Spanish (SP). This dataset is a collection of texts categorized by complexity level and annotated for complexity features, presented in Excel format (.xlsx). The corpora were compiled and annotated under the scope of the iRead4Skills.

Dataset 2 is derived from the previously released Dataset 1: Corpora by Level of Complexity for FR, PT, and SP (DOI: 10.5281/zenodo.10055909), which consists of written texts of various genres and complexity levels. A sample of texts from Dataset 1 was selected for classification and annotation, providing additional data and test sets for the complexity analysis systems in the three project languages.

Data Collection and Annotation Process

The classification and annotation tasks were carried out through a structured methodology:

Texts were distributed to Adult Learning (AL) and Vocational Education Training (VET) Centres, where trainers and students participated in classification tasks.
The classification was conducted via the Qualtrics platform, ensuring a standardized approach.
Participants assigned texts to one of four complexity levels:
- Very Easy (140 texts) – Easily understood by all.
- Easy (140 texts) – Understandable for those with less than 9 years of schooling.
- Plain (140 texts) – Readable at a 9th-grade level.
- More Complex (42 texts) – Challenging for individuals with a 9th-grade education.

For full details on the annotation process, data descriptions, and inter-annotator agreement, refer to the documentation available at Zenodo.