Using Natural Language Processing Tools on Individual Stories from First-year Students to Summarize Emotions, Sentiments, and Concerns of Transition from High School to College

Abstract

Prior to this, Dr. Satyanarayana was a Research Scientist at Microsoft in Seattle from 2006 to 2012, where he worked on several Big Data problems including Query Reformulation on Microsoft's search engine Bing. He holds a PhD in Computer Science from SUNY, with particular emphasis on Data Mining and Big data analytics. He is an author or co-author of over 25 peer reviewed journal and conference publications and co-authored a textbook-"Essential Aspects of Physical Design and Implementation of Relational Databases." He has four patents in the area of Search Engine research. He is also a recipient of the Math Olympiad Award, and is currently serving as Chair Elect of the ASEE (American Society of Engineering Education) Mid-Atlantic Conference. He also serves as an NSF (National Science Foundation) panelist.

Hey, let's take a selfie: insights of selfie defamiliarisation in the classroom

“…, 2019; Mite-Baidal et al. , 2018) in documents, sentences, or words (Sivakumar and Reddy, 2017) based on natural language processing (NLP) (Satyanarayana et al. , 2019) either by applying machine learning (supervised) or lexicon-based approach (unsupervised) (Mite-Baidal et al.…”

“…Combining SA and TA provides a deeper understanding of emotional experience from written feedback (Satyanarayana et al. , 2019), as certain tones could define emotions in textual expression (Kim and Ketenci, 2020).…”

“…The primary goal of SA is to classify students' opinions, feedback, review, and assessment (Chauhan et al, 2019) based on sentiments to identify attitudes and satisfaction (Zhou and Ye, 2020) with learning practices. SA is performed by extracting subjective information from unstructured text (Chauhan et al, 2019;Mite-Baidal et al, 2018) in documents, sentences, or words (Sivakumar and Reddy, 2017) based on natural language processing (NLP) (Satyanarayana et al, 2019) either by applying machine learning (supervised) or lexicon-based approach (unsupervised) (Mite-Baidal et al, 2018). The lexicon approach is sometimes called semantic orientation (SO), where a word's semantic intensity is used to determine the polarity and emotional valance (Awwad and Alpkocak, 2017).…”

“…(2) IBM Tone Analyzer Combining SA and TA provides a deeper understanding of emotional experience from written feedback (Satyanarayana et al, 2019), as certain tones could define emotions in textual expression (Kim and Ketenci, 2020). For that reason, we used the IBM TA, which is an application programming interface (API) developed by IBM Corporation (Al Marouf et al, 2019) that performs linguistic analysis (Sorvisto et al, 2018) of social and language tones (Mcgetrick, 2017) to detect emotions such as anger, fear, joy, and sadness and language tones namely analytic, confidence, and tentative (Mcgetrick, 2017).…”

Is Natural Language Processing Effective in Education Research? A case study in student perceptions of TA support

“…NLP approaches to sentiment analysis have achieved high accuracy levels, ranging from 75% to 99% when compared to traditional human approaches [17], [18]. More recently, sentiment analysis has been applied to a finer grained analysis of sentiment detecting tones of joy fear, sadness, anger, analytic, confident, and tentative in student generated stories of their lived experiences [19]. Student sentiment in feedback regarding their educational experiences is important, but it alone does not provide sufficient information to act on that feedback.…”