scikit-learn VS seqeval

I'd say, the first question you'd need to ask yourself is "Why do I want to do statistical tests" and "what kind of statistical tests do I want to do?". Most of them rely on a bunch of assumptions and just winging it will produce a number that will be reported and used but is terribly wrong. Funnily enough, scikit-learn does not directly give you p-values for this very reason and advise you to run the same regression in statsmodels.

Machine learning and statistical analysis? http://scikit-learn.org

Our next step is to create a new machine learning model based on this list. We’ll use Python’s excellent scikit learn framework to build our model. We’ll store our training data into two data frames: one for the set of features to train in and the second with the desired class labels. We’ll then split our dataset into 70% training data and 30% test data.

I can't tell you whether you'd like data science or machine learning, but I can tell you I took a class in it last year. It was an applied ML class targeting power systems engineers. ML is extremely statistics and probability heavy. I personally found the theory to be very dry, but the application to be rather enjoyable. We used sci-kit learn, which is an interesting Python package targeting academic data science and machine learning. https://scikit-learn.org/

scikit-learn for classical machine learning,

Scikit-learn - It is one of the most widely used frameworks for Python based Data science tasks. It features various classification, regression and clustering algorithms including support vector machines, random forests, gradient boosting, k-means and DBSCAN, and is designed to interoperate with the Python numerical and scientific libraries NumPy and SciPy. Link - https://scikit-learn.org/

Hi all,

Currently some works is being done to improve computational primitives of scikit-learn to enhance its overhaul performances natively.

You can have a look at this exploratory PR: https://github.com/scikit-learn/scikit-learn/pull/20254

This other PR is a clear revamp of this previous one:

Just to clarify, scikit-learn 1.0 has not been released yet. The latest tag in the github repo is 1.0.rc2

https://github.com/scikit-learn/scikit-learn/releases/tag/1....

Website: https://scikit-learn.org/ Github Repository: https://github.com/scikit-learn/scikit-learn Developed By: SkLearn.org Primary Purpose: Predictive Data Analysis and Data Modeling

. The standard way to evaluate NER (or any other sequence labelling problem) is to use the conlleval script (https://www.clips.uantwerpen.be/conll2000/chunking/output.html) or through the seqeval package in python (https://github.com/chakki-works/seqeval) . Either way, you need a list of predicted labels and a list of gold labels (see the code example in the link, it should be trivial to converse your output to the same data format).