Develop a ml classification model by Filinchet

FAQ

What do you need from me to start?

A sample dataset (CSV/Excel) and a brief description of the reaction systems and target classes.

What problem do you solve?

I build a machine-learning classifier to distinguish reaction types/systems from experimental variables (e.g., P, T, feed, WHSV, conversions).

What deliverables do I receive?

Trained model, cleaned notebook/script, performance report (precision/recall/F1), confusion matrix, and brief recommendations.

Which algorithms do you use?

Tree ensembles (Random Forest/XGBoost) by default; I can also test SVM/LogReg on request.

Can you handle imbalanced classes or small datasets?

Yes—using techniques like class weights, SMOTE/ADASYN, and careful cross-validation.

What if my data has missing values or noisy columns?

I include light preprocessing: type fixes, imputation, and feature selection as needed.

Do you explain the model’s decisions?

Yes—feature importance and optional SHAP plots to show what drives predictions.

Is my data confidential?

Absolutely—your data stays private; I can sign an NDA if required.

I will develop a ml classification model