Proprietary Data Scoring Algorithm Development

Job ID: 38130083

Budget: $30 – $250 USD

Request details
FOR EXPERTS WITH EXTENSIVE EXPERIENCE IN THE FIELD ONLY

Objective: Scoring Algorithm for evaluating potential investments

Data Source: 200+ proprietary historical records with up to 90 parameters

Mentoring requested (to be worked with tech project lead and data engineer):
• Data-Preprocessing (missing values, normalization, outliers)
• Implementation methods for feature selection (e.g. correlation Analysis, feature importance, etc.)
• Identify and run different algorithms (e.g., logistic regression, decision trees, random forests, gradient boosting, neural networks)
• Details and rationale behind algorithms selection
• Hyperparameter tuning and optimization
• Cross-validation
• Performance metrics (e.g., accuracy, precision, recall, F1 score, ROC-AUC)
• Score Calculation and thresholds
• Deployment and Monitoring

Tools available:
• In use: Microsoft Fabric and Excel (VBA)
• Potential tools: Python, R

Timeframe: 1 month