- Published on
Data transformation is key to effective machine learning. This post explores essential techniques in dealing with numerical data: normalization (linear, Z-score, log, clipping) and binning along with their pros and cons.
Machine-learning
All Posts
- computer-science (10)
- python (9)
machine-learning (5)
- project (3)
- c (2)
- dynamic-programming (2)
- text-classification (2)
- data-processing (2)
- data-science (2)
- docker (2)
- package (2)
- mutability (1)
- variables (1)
- data-types (1)
- bioinformatics (1)
- bugs (1)
- linear-alegbra (1)
- probability (1)
- naive-bayes (1)
- rasa (1)
- chatbot (1)
- nlp (1)
- encoding (1)
- software-engineering (1)
- angular (1)
- golang (1)
- mongodb (1)
- gin (1)
- agile (1)
- github (1)
- keras (1)
- open-source (1)
- build (1)
- deep-learning (1)
- linguistics (1)
- classification (1)
- cnn (1)
- neural-networks (1)
- computer-vision (1)
- demo (1)
- generative-ai (1)
- elo (1)
- games (1)
- cicd (1)
- feature-engineering (1)
- normalization (1)
- cryptography (1)
- security (1)
- tool (1)
- deployment (1)
- aws (1)
- devops (1)
- fastapi (1)
- Published on
Tutorial style Jupyter notebook on how to classify hindi (devanagari script) digits using Convolutional Neural Networks.- Published on
Gyan is a chatbot built using the open source framework Rasa, and can answer frequently asked questions about AI, Machine Learning and Data Science.- Published on
There are two common ways to encode categorical features. We can use Ordinal Encoding or we can use One Hot Encoding. Let's explore when you should use Ordinal encoding and when to use One hot encoding in this post.- Published on
I explore 4 methods used to classify text using the bag-of-words approach, we'll see the code and math required, and then I'll leave you with improvements we could make to these techniques in order to improve accuracy.