Weekly Progress IS
Week 1 – 4
In this week, we are starting to learn about Information System. For the first 5 weeks, we were being taught by sir ford. Other than that, we were also being taught by sir Andreas Kurniawan. We were being asked to find a topic for our final project but, we don’t know what to use. We are still blank on what we are going to do.
Later on, on week 2, my friend David tells us to just take stock prediction because we don’t know what we are going to pick for our project and Sir Andreas Kurniawan also keep asking us to get our topic. So we just stick with the stock prediction in the second week.
From week 3 – 4 we actually didn’t do anything yet for the final project and only so do some simple research on stock prediction algorithm.
Week 5 – 13
In week 5 we do some are still doing some searching on what algorithm that we are going to choose. But we are starting to feel confused because we don’t know some stock price meaning. In stock price, there are open, close, High, low, and volume for the basic data but we actually still don’t know what it is so we start to do some research on it.
In week 6, we get to know more about the stock but there is still another price that shows up such as adj close price, adj open, etc. We are still confused about this so we do another research on it. After that, we are still searching more on how to predict the stock and there is a lot of ways to predict the stock. And some simple logic on how to predict the price.
In week 7, we start to feel that knowing how to predict the stock price like seeing the trend and more don’t really matter in making machine learning because machine learning for stock prediction use algorithms and we will use a library to make it instead of making it from scratch. Because of that, we start to check the algorithm that we can use again
In week 8, we search and read a lot on medium.com, towarddatascience.com, and many more article about stock prediction. In the end, we choose to use random forest algorithms but we still not making it in code because we thought that it is still too hard for us as we still don’t really understand what it is and just randomly choosing random forest algorithm.
In week 9, after more searching, we found this one article on https://towardsdatascience.com/https-medium-com-vishalmorde-xgboost-algorithm-long-she-may-rein-edd9f99be63d
And in this article, it shows this image
Because of it, we feel that we need to change our algorithm to XGBoost as it is better. So we start to find more documentation on XGBoost.
In week 10, we try to code the machine learning for XGBoost but it is very hard and some of the documentation that we got is very weird in our eye and feels so hard. At the end, we just spend our time to try to code it.
In week 11, we feel that we got no hope in XGBoost and the deadline for it is nearing. Because of that, we decided to try other simpler algorithm that might work and was easier to code. That why we try to use Linear Regression and SVM (Support Vector Machine). This algorithm is much easier to code but the result that we get is kinda weird because the graph shows that the predicted value and the real value are very similar. Only all the data was being shifted up. That’s why we feel that we fail again and we were stuck once again.
In week 12, as the presentation is very near, we asked our lecturer about our project and asked Mrs. Nurul to guide us in the project. From then on, we found out that our project was already running just fine, only that we are using bad data so that the data seem bad.