Improving Multi-Class Classification with Stacking Ensembles and Feature Engineering: Need Insights

Hi everyone,

I am working on a machine learning task involving a multi-class classification problem with tabular, imbalanced data (no time series or categorical variables).

The goal is to predict class probabilities for a test set (150,000 rows x 9 classes) using models trained on the provided training data. To achieve lower log loss scores, I am exploring a multi-layered approach with stacking ensembles.

The first layer generates meta-features from diverse models (e.g., Random Forest, Extra Trees, KNN, etc.), while the second layer combines these predictions using techniques like LightGBM, SVM, or neural networks.

I am also experimenting with feature engineering (e.g., clustering, distance metrics, and embedding-based methods like UMAP and t-SNE), and advanced optimization techniques like Bayesian search for hyperparameters. Given the data imbalance, I am considering sampling techniques or class-weight adjustments.

Any suggestions or insights to refine this pipeline and improve model performance would be greatly appreciated.

 

2 Comments
 

It looks like this may be out of my ability to answer... maybe some of the links below might help?

  • For insights on machine learning and ensemble methods, you might find the discussions in the WSO forums on topics like "Machine Learning Taking over HF research analyst roles in near future?" helpful. It touches on the importance of data depth and variety for ML performance.
  • You can explore advanced optimization techniques and feature engineering strategies in threads like "Will robots replace your consulting or financial career?" which discusses AI's ability to handle large datasets with nuanced frameworks.

If you're looking for more specific advice, consider diving into machine learning-focused communities or resources.

I'm an AI bot trained on the most helpful WSO content across 17+ years.
 
Most Helpful

Optio est nihil ut animi est. Quia sed natus consequuntur sed in. Deserunt dolorem earum ut velit repellendus unde sequi fugiat. Eos qui reiciendis ea blanditiis.

Reiciendis inventore molestias voluptatem cumque aliquam. Sit sint et non velit laboriosam nihil. Non nulla in dignissimos iure et. Quisquam quibusdam rerum ab est eligendi quibusdam.

Career Advancement Opportunities

June 2026 Hedge Fund

  • Point72 99.0%
  • D.E. Shaw 98.1%
  • Citadel Investment Group 97.1%
  • AQR Capital Management 96.1%
  • Magnetar Capital 95.1%

Overall Employee Satisfaction

June 2026 Hedge Fund

  • Magnetar Capital 99.0%
  • D.E. Shaw 98.0%
  • Blackstone Group 97.0%
  • Citadel Investment Group 96.0%
  • Millennium Partners 95.0%

Professional Growth Opportunities

June 2026 Hedge Fund

  • AQR Capital Management 99.0%
  • Point72 98.1%
  • D.E. Shaw 97.1%
  • Citadel Investment Group 96.2%
  • Magnetar Capital 95.2%

Total Avg Compensation

June 2026 Hedge Fund

  • Portfolio Manager (9) $1,648
  • Vice President (27) $464
  • Director/MD (12) $423
  • NA (9) $320
  • Engineer/Quant (86) $288
  • 3rd+ Year Associate (26) $284
  • Manager (4) $282
  • 2nd Year Associate (32) $253
  • 1st Year Associate (76) $192
  • Analysts (242) $181
  • Intern/Summer Associate (28) $146
  • Junior Trader (5) $102
  • Intern/Summer Analyst (282) $96
notes
16 IB Interviews Notes

“... there’s no excuse to not take advantage of the resources out there available to you. Best value for your $ are the...”

Leaderboard

1
redever's picture
redever
99.2
2
Secyh62's picture
Secyh62
99.0
3
kanon's picture
kanon
99.0
4
BankonBanking's picture
BankonBanking
99.0
5
CompBanker's picture
CompBanker
98.9
6
GameTheory's picture
GameTheory
98.9
7
DrApeman's picture
DrApeman
98.9
8
dosk17's picture
dosk17
98.9
9
Betsy Massar's picture
Betsy Massar
98.9
10
Mimbs's picture
Mimbs
98.8
success
From 10 rejections to 1 dream investment banking internship

“... I believe it was the single biggest reason why I ended up with an offer...”