Machine Learning, Deep Learning
My explanation of the core difference between AI algorithms
Click here to get to know me, my projects, and my latest articles.
One of the first problems a beginner data scientist faces is understanding the difference between Deep Learning and Machine Learning. In general, Machine Learning englobes the entire world of AI, while Deep Learning is a sub-category of Machine Learning. However, this does not give you any valuable insight. In this article, I would like to dig deeper into the structural differences between Machine Learning and Deep Learning algorithms, so that you can understand what is their key difference and how to spot them.
Why understanding the difference?
If you are aspiring…
Deep Learning
Detailed explanation and algorithms, improvement of my previous code on the matter. Full code available at my repo.
Click here to get to know me, my projects, and my latest articles.
In 2020 I already published a similar article on how to perform digit classification. Although the code and the results are almost identical (is the same problem, after all), this year, after learning much more about deep learning architecture, I learned more about my code and improved, even more, introducing new functions and, hopefully, being able to explain all components of the notebook.
Most difficult concepts
When performing image classification, there are a few meta-steps that are common to all similar models when using CNN, but may cause some confusion…
Videogame
Seems unbelievable, but crackers seem to be having a tough time with this game
Click here to get to know me, my projects, and my latest articles.
Disclaimer
This article gives an update on the current state of the piracy industry, I have not included any link to any pirated/illegal content. Regarding the links to CrackWatch, according to their website: “CrackWatch.com does NOT host any pirated or illegal content, does NOT provide links to any illegal files such as torrents and is NOT associated with any scene group.”
93 days counting
After cracking Big Titles such as CyberPunk 2077, Death Stranding, and the long-awaited Red Dead Redemption, it has been 93 days since the release of the Ubisoft…
Quantitative Finance, Data Visualization, Programming
Procedurally generating synthetic stock data for simulation purposes. Full code available at my repo.
Click here to get to know me, my projects, and my latest articles.
One of the toughest parts of technical analysis is finding the data. Downloading historical stock data consistently is not as easy as it sounds: you need an API key from a company that provides the data, then data has to be downloaded, parsed, cleaned, and only then you can have some data to work on. In fact, simply using two or three stocks as a sample is interesting, but is a bit restrictive when you are experimenting with the full repertoire of algorithms in quantitative finance.
As…
Finance
A clear explanation of the GameStop issue in the Finance world
We are continuously publishing reports consisting of news and quantitative market analysis. For additional information about us and our team of analysts visit Yeetum.
***There are several interpretations of this story: some side for the Brokerages which have been taking extreme actions, some others from the Reddit community which is condemning the brokerages. We have been following only public statements, we are not picking any side.
In the last few days, GameStop, a publicly-traded videogame retailer, has been at the center of attention because of brokerage intervention on its stock. …
Data Visualization
Compressing 21 dimensions of numerical data into 3, then graphing the results in 3D
The data has been downloaded from this website, which is an open-source dataset of movie ratings and tags. I will edit the data in the dataset for visualization purposes, encoding all tags as one_hot encoding, and placing them into a multidimensional space.
Because a space of more than 3 dimensions cannot be graphed by using an image, I will need to use PCA techniques to compress the data and visualize it in a 3D space using Plotly.
Download the data
Because data visualization is a GPU intensive task, with several limitations in a 3D space, I can simply use a 2MB dataset. If…
Data Visualization
Can I tell how good is my data and how good can my model be just by looking at the data? Full code available at my repo.
Creating a wine classifier is one of the easiest experiments that is proposed to beginners during their first Machine Learning project. In this tutorial, I already teach in detail how to build one, if you are interested in the model feel free to study it and experiment with the code I provide with it.
However, this article will focus on data visualization. Specifically, how can I estimate how well I can build a model using ML from the data, and, if necessary, if deep learning may be required for a more flexible inference function.
Mathematics
Solving math riddles with Monte-Carlo simulations. Full code available at my repo.
One of the most interesting math problems I have encountered this year is in the game dynamics of Clash Royale, a game that hides many challenging but creative mathematical riddles (some of which I already tackled, if you are interested read how I applied Clustering techniques for Clash Royale decks).
The math riddle
The game allows players to participate in challenges. Every time the player enters a match has to play with another player and can either win or lose (draw is not made possible by the new game dynamics). When the player enters the challenge has 3 lives, and can win as…
Cryptocurrencies
The latest news from a quantitative finance perspective on cryptocurrencies
The analysis has been performed using python language, with an extended use of matplotlib, seaborn, plotly, pyportfolioopt, and sklearn.
We are publishing a weekly report on our market analysis. For additional information about us and our team of analysts visit Yeetum.
Qualitative Report
The Bitcoin hype seems to have reached its peak, at least for this cycle. After the cycle has corrected, investors will average down on Bitcoin's future prosperity.
What we are witnessing right now is just another step in the Bitcoin life cycle. The market volume for Bitcoin has increased substantially in the past few years, allowing investors to perform…
Finance
Version 2.0, with some changes to my previous code.
Alpha Vantage is one of the possible online tools that you can use to download Stock historical data using code. This website provides you with a personal API key that you can use to download a limited amount of data (up to 500 free requests per day, which is not bad).
Feel free to check my previous article with the deprecated version of my code and an in-depth analysis of the reasons that I choose alpha vantage over many other API finance providers.
Why using code to download stocks
If you had to download a single stock you can easily use the yahoo finance website. However…
