NATURAL LANGUAGE PROCESSING AS A METHOD FOR EVALUATION OF FACTORS INFLUENCING SMILE ATTRACTIVENESS

Abstract

Introduction: A drastic increase in the number of published medical papers per year is observed. This makes the identification, analysis and categorization of significant studies a difficult task. Natural (human) Language Processing and text mining are methods, part of the scientific branch computer linguistics that transfer the informational overload from a human to a computer. It enables easier processing and analysis of large volumes of unstructured textual data. Purpose: The current study aims to familiarize researchers working in the field of dentistry with the capabilities of NLP and TM for a quick and concise analysis of large volumes of unstructured textual information and identification of dependencies between different factors important for a given subject. Materials and Methods: To demonstrate the capabilities of text mining, an important topic in the field of dentistry was chosen – factors influencing the esthetics of a smile. The analysis was carried out with “R”- a computer language for statistical processing. A literature search was conducted in the “PubMed” database with key-words – “dental, esthetic and factor”. The resulting abstracts were saved as a local copy, imported and processed. Word frequencies and associations between different terms were analyzed. Results and discussion: Weak to moderate correlation was established between the significant, most frequent terms in the text - “esthetics, „smile“, „arc“, „buccal“, „gingival”, “lip” and “midline”./0.1<r<0.45/ Word combinations and frequencies resulting from the analysis are in agreement with other reported findings. Conclusion: NLP and text mining are valuable tools which decrease the time necessary for analysis of large volumes of data. The results can aid further research with increased accuracy.

Keywords

• ext mining
• natural language processing
• smile
• factors
• esthetics