Hi Xi, great job on your poster! First of all, the layout of your poster is very good, especially with the great balance of descriptions, equations and graphs. With the use of bullet points and words in bold, audience can quickly obtain the information about your model specifications. I wonder if there might be a better way to introduce the six classical models? I really enjoyed the clear explanations that you delivered for each model in your presentation, and I wonder would it be possible to just add a bit more explanations on your poster?
The graph that compares classical models by mean absolute deviation across conditions is really helpful for me to understand why you would like to propose new models which somehow combine the specification of classical models. It is pretty impressive that your new model 2 has the performance as good as ITCH model, which means that the term (t2-t1)/t* in ITCH model could be substituted by (d2-d1). This is an excellent extension for existing classical model (ITCH) and it probably has better overall performance based on 5 conditions. Additionally, I think the section of conclusion is very precise with great summaries on the key findings.
Overall, I really like your poster design and good luck with your final paper!
Hi Xi, great job on your poster! First of all, the layout of your poster is very good, especially with the great balance of descriptions, equations and graphs. With the use of bullet points and words in bold, audience can quickly obtain the information about your model specifications. I wonder if there might be a better way to introduce the six classical models? I really enjoyed the clear explanations that you delivered for each model in your presentation, and I wonder would it be possible to just add a bit more explanations on your poster?
The graph that compares classical models by mean absolute deviation across conditions is really helpful for me to understand why you would like to propose new models which somehow combine the specification of classical models. It is pretty impressive that your new model 2 has the performance as good as ITCH model, which means that the term (t2-t1)/t* in ITCH model could be substituted by (d2-d1). This is an excellent extension for existing classical model (ITCH) and it probably has better overall performance based on 5 conditions. Additionally, I think the section of conclusion is very precise with great summaries on the key findings.
Overall, I really like your poster design and good luck with your final paper!