Scientific Text Mining Research Articles

Critical thinking and continuous improvement: a scientific text mining approach

This work aims to propose and argue a new antecedent (critical thinking: CT) of the hard and soft dimensions of continuous improvement (CI) using a text mining perspective. The study employs a proposed methodological framework for scientific text mining (nine steps). The sample consists of 2,213 abstracts of articles that address the interest constructs in this study from the title. Three types of variables are used: investigation, reference, and control. The article presents theoretical and empirical evidence for an original model conformed by six propositions that relate CT, CI, dynamic capabilities that potentiate CI, and the context (academic, business), which place CT as a trigger for CI. The contrast of this model provides useful inputs for teachers, researchers, professionals, students, and leaders so that they can better motivate and argue the need to strengthen CT when CI is taught and applied (classically limited to the use of tools). This article provides two contributions. The first is related to the effect of CT on the development (process and result) of CI. The second consists of providing a framework to contrast structural hypotheses from a scientific text mining approach.

A neural approach for detecting inline mathematical expressions from scientific documents

AbstractScientific documents generally contain multiple mathematical expressions in them. Detecting inline mathematical expressions are one of the most important and challenging tasks in scientific text mining. Recent works that detect inline mathematical expressions in scientific documents have looked at the problem from an image processing perspective. There is little work that has targeted the problem from NLP perspective. Towards this, we define a few features and applied Conditional Random Fields (CRF) to detect inline mathematical expressions in scientific documents. Apart from this feature based approach, we also propose a hybrid algorithm that combines Bidirectional Long Short Term Memory networks (Bi‐LSTM) and feature‐based approach for this task. Experimental results suggest that this proposed hybrid method outperforms several baselines in the literature and also individual methods in the hybrid approach.