Ammu Bharat Ram – Medium

Ammu Bharat Ram

Pinned

Ammu Bharat Ram

How to Build a Secure Production-Grade Web App for Banking Using Python

In the competitive world of banking, security, scalability, and efficient management of applications are paramount. When it comes to…

Jun 23

Jun 23

Pinned

Ammu Bharat Ram

Fuzzy Name Matching in Python: Comparing Customer Names with Watchlist Entities as part of…

Introduction

May 2, 2023

May 2, 2023

Ammu Bharat Ram

Publishing Your Personal Website for free Using GitHub Pages: A Step-by-Step Guide

Introduction: In today’s ever-connected world, establishing a robust online presence is paramount for data scientists like us. Enter GitHub…

Jun 30, 2023

Jun 30, 2023

Ammu Bharat Ram

dependency of categorical encdoing on current target value (like in Ridit) , then its leakage.

Apr 12, 2021

Apr 12, 2021

Ammu Bharat Ram

model.get_feature_importance(Pool(X_test, label=y_test,cat_features=categorical_features_indices),

type="ShapValues")

Mar 22, 2021

Mar 22, 2021

Ammu Bharat Ram

to reduce time for predicting

Mar 22, 2021

Mar 22, 2021

Ammu Bharat Ram

Hence Cat is more suitable for production environment

Mar 22, 2021

Mar 22, 2021

Ammu Bharat Ram

Cat:

If onehot size > one_hot_max_size, Cat uses response coding

Mar 22, 2021

Mar 22, 2021

Ammu Bharat Ram

if time not there in data, it will thus require to create artificial time in Cat

Mar 22, 2021

Mar 22, 2021

Ammu Bharat Ram

only for catboost, similar to ridit scoring

Mar 22, 2021

Mar 22, 2021

Ammu Bharat Ram

Ammu Bharat Ram

Head of Data Science at Complidata. Expert at Python, MLOps, LLM, ML web apps using Python. Master's in Data Science from KU Leuven.

Following

Help
Status
About
Careers
Press
Blog
Privacy
Terms
Text to speech
Teams