Analytics Recipe Cookbook - Summaries

 View Only

Evaluate your ML model for audience building with Precision & Recall

  • 1.  Evaluate your ML model for audience building with Precision & Recall

    Posted 11-19-2020 03:28 PM

    Introduction:  

    This recipe provides a quick guide as to why precision and recall are important metrics for marketers when dealing with a classification problem such as identifying key customers for building audiences.


    Analysis Overview:

    Marketers are increasingly using machine learning to predict user behavior. Some of their most common goals are: identifying high-value customers, finding lookalike (prospects) customers, providing relevant content, and building an overall positive customer experience. In data science, classification modeling is a great approach for identifying key customers. Classification modeling can help answer the question of whether customers fit in one group or another, which is great for segmenting and audience building. Examples of classification problems include

    • Which of my customers will purchase product A?
    • Which of my customers will subscribe to my service?
    • Which of my customers are likely to churn?

    Analysis Benefits:

    • Identify which group you customers fit in, which is great for segmenting and audience building
    • Tackling a classification problem in an imbalanced dataset 
    • Evaluation metrics and techniques to deal with imbalanced data

    DAA members, view the full recipe here.

    ------------------------------
    Shasnika DeMel
    Advanced Analyst - Data Science
    InfoTrust LLC (Corporate Account)
    Toronto ON
    6473959033
    ------------------------------
    2020 compensation survey now open