TFI Restaurant Revenue Kaggle Competition

Read my report HERE and R code is provided HERE (warning: messy but all-inclusive code) I recently entered into my first Kaggle competition with a friend while in my graduate (I'm undergrad) Data Mining class.  At first I thought this course was going to be super difficult and while the content is certainly overwhelming, I felt that most people were pretty much at the same level of understanding that I had.  The class was super interesting and although I could not get to know all the details for each chapter, they certainly opened my eyes to how amazing and vast […]

Continue reading →

Consumer sentiment analysis used for Finance??

I recently just finished a paper with my partner in my data mining course. We were assigned the topic of Naive Bayes classifier and we decided to apply them to sets of Amazon user reviews in a bunch of categories.  I've uploaded the paper on this blog in case you are interested in reading the details of the data mining. In summary, the main results and implications of our research were: It is possible to accurately classify consumer sentiments through analyzing and identifying key words in the review. We can get more accurate predictions and meaningful data by examining reviews […]

Continue reading →

New Features, New functions, Lower errors

This will probably be the last post I have regarding PAM.  All the project files can be found on my github page.   I don't think I'll be working with her anymore as I don't think theres anymore methodologies to drastically improve her error with JUST price action.  It's not worth my time and I need to move on.  I introduced some new features such as Gap, Previous ratios, etc.  They did lower the errors again. I have come a long way from validation errors of approx 0.6938 to 0.4955.  Out-of-sample pairs forecasting is still inconclusive as the model did […]

Continue reading →

Implementing Trader into PAM

As a follow up from the previous post, I made quite a few errors designing Pam's feature input and have now fixed them.  In addition I introduced Pam two new features: Body Length ratio between current candle and last candle Body Length to Wick Ratio of current candle I retrained Pam with a very high lambda and less amount of nodes () .  The model overall had a better Validation and Testing fit.  I also skipped out on adding polynomial features for now and only relational (ex: ) features. I decided to bring it to the test set of EUR/USD […]

Continue reading →

Meet Pam

Project files repo Pamela, you can call her Pam, recently started Forex trading.  She is a complete beginner but she is extremely hard working so I believe she can be profitable soon. Pam, unlike most forex trader, is not human.  PAM's name derives from Price Action Machine, an artificial neural network that currently learns from candlestick patterns on currency pairs.  She has a simple neural network structure as such: With number of independent (feature) variables and number of nodes in the hidden layer.  The connections between input to hidden and to output consists of two matrix of weights.  We can call them […]

Continue reading →