Dataset: Annotated soybean market news articles.

This dataset involves a collection of soybean market news through web scraping from a Brazilian website. The news articles gathered span from January 2015 to June 2023 and have undergone a labeling process to categorize them as relevant or non-relevant. The news labeling process was conducted under the guidance of an agricultural economics expert, who collaborated with a group of nine individuals. Ten parameters were considered to assist participants in the labeling process. The dataset comprises approximately 11,000 news articles and serves as a valuable resource for researchers interested in exploring trends in the soybean market. Importantly, this dataset can be utilized for tasks such as classification and natural language processing. It provides insights into labeled soybean market news and supports open science initiatives, facilitating further analysis within the research community.

On the enrichment of time series with textual data for forecasting agricultural commodity prices.

Forecasting models in the financial market generally use quantitative time-series data. However, external factors can influence data in time-series, such as weather events, economic crises, and the foreign exchange market. This information is not explicit in the time-series and can influence the prediction of the variable values. Textual data can be a source of knowledge about external factors and is potentially helpful for time-series forecasting models. Some studies have presented text mining techniques to combine textual and time-series data. However, the existing representations have limitations, such as the curse of dimensionality and sparse data. This work investigates the finite use of domain-specific terms to investigate these problems by representing textual data with low dimensional space. We consider thirty-three keywords that are potentially important in the domain to enrich time-series using text mining techniques. Four regression models were applied to the representation proposed to predict the future daily price of corn and soybeans. The experimental setup considers a real market scenario, in which the daily sliding window strategy and step-forward forecast were used. The representation proposed has better accuracy in some forecasting scenarios. The results indicate that text data are a promising alternative for enriching time-series representations and reducing uncertainty forecasting models.•We show an approach to enriching time-series using domain-specific terms;•Representation proposed combines quantitative data with qualitative market factors;•Regression Models to learn a forecasting function from enriched time-series.

A network-based positive and unlabeled learning approach for fake news detection.

Fake news can rapidly spread through internet users and can deceive a large audience. Due to those characteristics, they can have a direct impact on political and economic events. Machine Learning approaches have been used to assist fake news identification. However, since the spectrum of real news is broad, hard to characterize, and expensive to label data due to the high update frequency, One-Class Learning (OCL) and Positive and Unlabeled Learning (PUL) emerge as an interesting approach for content-based fake news detection using a smaller set of labeled data than traditional machine learning techniques. In particular, network-based approaches are adequate for fake news detection since they allow incorporating information from different aspects of a publication to the problem modeling. In this paper, we propose a network-based approach based on Positive and Unlabeled Learning by Label Propagation (PU-LP), a one-class and transductive semi-supervised learning algorithm that performs classification by first identifying potential interest and non-interest documents into unlabeled data and then propagating labels to classify the remaining unlabeled documents. A label propagation approach is then employed to classify the remaining unlabeled documents. We assessed the performance of our proposal considering homogeneous (only documents) and heterogeneous (documents and terms) networks. Our comparative analysis considered four OCL algorithms extensively employed in One-Class text classification (k-Means, k-Nearest Neighbors Density-based, One-Class Support Vector Machine, and Dense Autoencoder), and another traditional PUL algorithm (Rocchio Support Vector Machine). The algorithms were evaluated in three news collections, considering balanced and extremely unbalanced scenarios. We used Bag-of-Words and Doc2Vec models to transform news into structured data. Results indicated that PU-LP approaches are more stable and achieve better results than other PUL and OCL approaches in most scenarios, performing similarly to semi-supervised binary algorithms. Also, the inclusion of terms in the news network activate better results, especially when news are distributed in the feature space considering veracity and subject. News representation using the Doc2Vec achieved better results than the Bag-of-Words model for both algorithms based on vector-space model and document similarity network.

Automation of Article Selection Process in Systematic Reviews Through Artificial Neural Network Modeling and Machine Learning: Protocol for an Article Selection Model.

BACKGROUND: A systematic review can be defined as a summary of the evidence found in the literature via a systematic search in the available scientific databases. One of the steps involved is article selection, which is typically a laborious task. Machine learning and artificial intelligence can be important tools in automating this step, thus aiding researchers. OBJECTIVE: The aim of this study is to create models based on an artificial neural network system to automate the article selection process in systematic reviews related to "Mindfulness and Health Promotion." METHODS: The study will be performed using Python programming software. The system will consist of six main steps: (1) data import, (2) exclusion of duplicates, (3) exclusion of non-articles, (4) article reading and model creation using artificial neural network, (5) comparison of the models, and (6) system sharing. We will choose the 10 most relevant systematic reviews published in the fields of "Mindfulness and Health Promotion" and "Orthopedics" (control group) to serve as a test of the effectiveness of the article selection. RESULTS: Data collection will begin in July 2021, with completion scheduled for December 2021, and final publication available in March 2022. CONCLUSIONS: An automated system with a modifiable sensitivity will be created to select scientific articles in systematic review that can be expanded to various fields. We will disseminate our results and models through the "Observatory of Evidence" in public health, an open and online platform that will assist researchers in systematic reviews. INTERNATIONAL REGISTERED REPORT IDENTIFIER (IRRID): PRR1-10.2196/26448.