https://coffepowered.github.io/tags/machine-learning/Recent content in Machine-Learning on AW — notes & thoughtsHugo -- gohugo.ioit-ITAndrea RuggeriniAndrea RuggeriniWed, 01 Mar 2023 21:28:43 -0500https://coffepowered.github.io/blogs/2022-sparse-notes/Wed, 01 Mar 2023 21:28:43 -0500https://coffepowered.github.io/blogs/2022-sparse-notes/<h2 id="data-science">Data science</h2> <ul> <li><code>march</code> <a href="https://bookdown.org/max/FES/">Effect of irrelevant features</a> - a nice experimental test</li> <li><code>march</code> <a href="https://www.coursera.org/learn/competitive-data-science/lecture/8o1Hc/matrix-factorizations">Matrix factorizaton recap in 6minutes</a></li> <li><code>march</code> <a href="https://www.kaggle.com/code/vprokopev/mean-likelihood-encodings-a-comprehensive-study/notebook">mean encodings: a comprehesive study</a>, check also <a href="https://github.com/scikit-learn-contrib/category_encoders">category_encoders sklearn library</a></li> <li><code>march</code> relationship between <a href="https://stats.stackexchange.com/questions/134282/relationship-between-svd-and-pca-how-to-use-svd-to-perform-pca">SVD and PCA</a> + a comprehensive <a href="https://arxiv.org/pdf/1404.1100.pdf">tutorial on arxiv</a></li> <li><code>march</code> Naive-Bayes classification from <a href="https://jakevdp.github.io/PythonDataScienceHandbook/05.05-naive-bayes.html">scratch</a></li> <li><a href="https://arxiv.org/abs/2203.05556">On Embeddings for Numerical Features in Tabular Deep Learning</a></li> <li><a href="https://cmry.github.io/notes/euclidean-v-cosine">Euclidean vs cosine distance</a></li> </ul> <h2 id="data-viz">Data viz</h2> <h2 id="complimentary">Complimentary</h2> <h2 id="temporary">Temporary</h2> <p>Staging area for links that will be forgotten soon:</p> <ul> <li><code>march</code> <a href="https://www.kaggle.com/competitions/h-and-m-personalized-fashion-recommendations/data">H&amp;M personalized fashion</a> reccomendation</li> <li><code>march</code> learning <a href="https://wattenberger.com/blog/d3">D3js</a> + <a href="https://www.linkedin.com/learning/d3-js-essential-training-for-data-scientists/what-you-need-to-know">essential training</a> + custom visuals into <a href="https://www.linkedin.com/learning/advanced-power-bi-custom-visuals-with-d3-js/intro-to-microsoft-power-bi?u=2057564">powerbi</a></li> </ul>https://coffepowered.github.io/blogs/google-stock-split/Wed, 02 Feb 2022 21:28:43 -0500https://coffepowered.github.io/blogs/google-stock-split/<p>Google announced Tuesday its results for Q4 2021 <em>and</em> that they plan to split shares at 20 for 1 this Tuesday. Basically, if you are a shareholder, for each 1 share of Google at (about) 3000$ dollars you will have (at the ex-date) 20 shares at 150$.</p> <p>Nothing should really change, but there might be some financial effects, for instance, due to the <a href="https://www.barrons.com/articles/alphabet-stock-split-51643842562">inclusion of the stock in new indices</a> or increased market liquidity.</p>