DS821: News and Market Sentiment Analytics
Entry requirements
Academic preconditions
It is assumed that students are familiar with the following concepts:
- Basic classification and regression techniques.
- The concepts of train-, validation- and test-split of data.
- Sampling techniques.
- Implementation of mentioned techniques in a programming language.
The courses below, which are all part off the profiles in Data Science all gives a high degree of understanding of the above concepts.
- Economics and Business Administration: DS810: Data-driven Decision Making, DS804: Datamining and machine learning, DS831: Programming for Data Science, DS805: Multivariate Statistical Analysis
- Health Data: DS804: Datamining and machine learning, DS831: Programming for Data Science, DS805: Multivariate Statistical Analysis, DS812: Introduction to Basal Biostatistical Terms and Regression
- Environmental Data Science: DS804: Datamining and machine learning, DS831: Programming for Data Science, DS805: Multivariate Statistical
- Human informatics: DS804: Datamining and machine learning, DS831: Programming for Data Science, DS805: Multivariate Statistical Analysis, DS807: Applied machine learning
- ICT Systems: DS820: Discrete methods for Data Science, DS830: Introduction to programming, DS804: Datamining and machine learning, DS807: Applied machine learning, DS805: Multivariate Statistical Analysis
Course introduction
The course provides a foundation for applying the knowledge, competences, and skills, which are acquired in previous machine learning and regressions courses particularly addressing issues related to text analysis to facilitate a basis for news and market sentiment analysis.
In relation to the competence profile of the degree it is the explicit focus of the course to:
- give competences to understand and handle large amounts of text data
- give skills to analyse and reflect on transforming raw text data into useful input for classification and/or regression models
- give knowledge on how to understand and frame questions related to sentiment analysis
- give skills to formulate and identify possible consequences of chosen procedures when handling text data
- give the students a basic toolbox for text analysis in order to establish a basis for future work within NLP and/or sentiment analysis in a broad sense
- give the students knowledge on available NLP and sentiment tools
- give skills to adopt machine learning in combination with text data to achieve useful and presentable results
Expected learning outcome
The learning objective of the course is that the student demonstrates the ability:
- To write simple programs that manipulate and analyse language data.
- To understand the key concepts of NLP and linguistics used to describe and analyse language.
- To be able to evaluate performance of NLP techniques based on data.
Content
The following main topics are contained in the course:
- Applications of large language models and transformers
- Leveraging financial news as data
- Basics of NLP
- Deep learning, embeddings and feature extraction
- Text and sentiment classification
- Text as data
Literature
Examination regulations
Exam element a)
Timing
Tests
Written take home exam
EKA
Assessment
Grading
Identification
Language
Duration
Examination aids
To be announced during the course
ECTS value
Indicative number of lessons
Teaching Method
At the faculty of science, teaching is organized after the three-phase model ie. intro, training and study phase.
- Intro phase (lectures): 6 hours
- Training phase: 18 hours, including exercises 15 hours.
Activities during the study phase:
- Preperation of lectures
- Preparation of exercises
- Casework (news and/or market analysis)
Teacher responsible
Name | Department | |
---|---|---|
Christian Møller Dahl | cmd@sam.sdu.dk | Econometrics and Data Science |
Christian Vedel Sørensen | christian-vs@sam.sdu.dk | Økonomisk institut |