Today I studied preprocessing for basic learning models. - Dropping and imputing missing values (dropna, SimpleImputer) - Handling categorical data (mapping ordinal features, one-hot encoding nominal features and its downsides) - Splitting and scaling datasets Next time I'll put it into practice with the Titanic CSV! I'd like to try Kaggle too. #ML
// notes
Daily Notes
A log of short thoughts and notes.
Today I studied the basic learning models, perceptron and ADALINE, plus KNN, decision trees, random forests, and logistic regression. I skipped over a fair amount of the math, so next time I'll go deeper. Random forests are nice. #ML
I used to think React was complicated and just didn't get it. But then I learned that back in the jQuery days there was this problem where you couldn't read anyone else's code at all, and React solved it at the structural level. Suddenly it all clicked why we write things the way we do now. Tech gets a lot friendlier once you learn its history. (I still don't get it at all 😭) https://blog.shinonome.io/react-and-frontend-philosophy/ #WebDev
Reactの発明とフロントエンド思想の変遷
フロントエンド技術の歴史は、「何が辛かったか」の連鎖だ。その中で、Reactはかつてフロントエンド開発の景色を一変させた。 しかし今、htmxやHotwireといった新たな選択肢が注目を集める中で、「Reactはもう不要だ」という声が増えている。 はたしてReactは本当に不要なのだろうか。それとも、私たちはReactが何を解決したのかを忘れてしまっただけなのだろうか。 本記事では、Web黎明期から現代に至るフロントエンド技術の変遷を辿り直し、各技術の功罪と、技術選定において本当に問うべきことを考える。
blog.shinonome.io
Had my Coming of Age ceremony today. Got to see everyone after a long while, and the reunion was fun. #Life