Our primary goal is to provide a tool for faculty and research students to explore the research topics in Finance.
FinanceTopic could be your first destination to conduct literature review, select a dissertation or research topic, find potential reviewers and collaborators for your project, and so on.
We have identified 300 research topics in Finance at the finest granularity level; for each topic, you can view all relevant published articles in top journals and also prominent researchers working on that topic.
Steps to take:
Our secondary goal is to provide a profile page for every published researcher in Finance. You can browse each researcher's research interests, coauthor network, and journal publications.
TopOur topic search function is built on the outputs of topic models.
A topic is defined as a list of words (more accurately, a distribution over these words);
the order of appearance in the list for each word is determined by its importance to the topic. For example, the "IS Security" topic has words about security with high probability,
such as security, risk, attack, and disclosure; these words would appear at the beginning of the word list for this topic.
An article (or an abstract) could contain multiple topics with different relevancy.
The most relevant topic in an article would have the highest probability; irrelevant topics would have a probability of zero or close to zero.
Once you find a topic that is relevant to the keywords you specify, we will show all the articles that are most relevant to that topic.
Searching keywords in titles/abstracts requires that the keywords have to appear in the text of a paper,
which often misses the papers that are still relevant to the topic you are searching for but do not explicitly mention the particular keywords.
We apply Correlated Topic Model (CTM) and Latent Dirichlet Allocation (LDA) to identify the research topics in Finance at different levels of granularity. One input parameter for topic models is the number of topics to be generated. By evaluating the held out accuracy and the interpretability of generated topics, we find an optimal number of topics to produce the topic model results. For Finance, we find that LDA-300 offers the best model fit.
TopTopic models are built on the abstracts of research articles/notes published in the top Finance journals in the UTD list. Currently, we include articles up to 2014.
TopServer logging has been enabled on our site. By studying how you use FinanceTopic, we will keep improving our services and also share with you insights into what is trending among the Finance community.
Top