Unsupervised learning ML model to detect outliers in the data. There are 4 behavioral variables and 37 contextual attributes in the data. Many categorical (string vbs) and numerical vbs need to be normalized. You should also measure the performance of your algorithms by certain evaluation metrics and visualization.

Question

Sigiloso · Accepted Answer

I used Isolation Forest, best known in the industry combined with PCA to reduce the dimension of the contextual features. I also applied normalization to the data initially. I used Cross-Scoring for the evaluation of the algorithm.

Adyen

Pergunta de entrevista da empresa Adyen

Resposta da entrevista

Empresas seguidas

Buscas de vagas