feature engineering

"Everyone online is screaming about Agentic AI, LLM wrappers, and prompting techniques. Meanwhile, I'm just sitting here trying to wrap my head around basic regression models and proper feature engineering."

"I built one mini projects today. - Titanic Survival Predictor I learnt : - Handling real world dataset - Data cleaning - Converting text to numbers ( Encoding)"

"Python’s itertools module is a standard library toolkit that most data scientists know exists but rarely reach for when building features."

"Stop burning your GPU (and your time) for no reason. Start simple, then earn the right to get complex."

"However, in horse racing data, if you 'put raw data as is', even if the apparent AUC increases, ROI may decrease."

"Specifically, you can leverage pre-trained LLMs from providers like Groq (for example, models from the Llama family) to undertake data transformation and preprocessing tasks, including turning unstructured data like text into fully structured, tabular data that can be used to fuel predictive machine learning models."

"This is the process of converting what humans see in horse racing, like analyzing horse performance or jockey skills, into 'numerical values' that AI can understand."

"This article explores the issue of data leakage, and a solution is introduced: a "fingerprint file system" that saves the order and data types of columns at the time of learning."

"By using time-oriented feature creation (1st and 2nd order differences, etc.) and metadata embedding, and using Transformer and BiLSTM as backbones, good results were obtained."

"This blog post implements the entire process from importing CSV data to feature engineering, model training, registration in the Model Registry, and inference of test data."

"Clustering means assigning data points to groups based on their similarity, in other words, the idea of "birds of a feather flock together.""

"This article details the use of LightGBM, a decision-tree-based machine learning model."

"The article suggests exploring the idea of learning local structures (patches) with k-means and then using them as features, which is "clustering as feature engineering.""

"In this article, we'll build a system to predict housing prices using machine learning, using the open data "California Housing Price Dataset.""

"Results show that AI can be substituted."

"The article discusses adding and processing specific features from past race data."

"On a novel in-house dataset, our approach achieves 38% and 150% improvement in the evaluation metric over manually crafted and unplanned workflows respectively."

"The article mentions using Gemini for implementation."

"The article is the 11th in a series where a programming beginner uses ChatGPT to build a horse racing prediction AI to learn about Generative AI and programming."

"Lambdo is presented on Hacker News, suggesting it's likely a new tool or project."

"The article's source is Hacker News."