Machine-learning project

Demand_Side_Platform_Click_Through_Rate

Table of contents

• General info
• Problem statement
• Setup
• Pre and Post-Processing
• Expected Output

General info

This document provides instructions on how to build an api that can integrate a Click Prediction Service into various solutions by making a request to the service's HTTP application endpoint. The documentation is intended for developers/users who want to write applications that can interact with the Click Prediction Service API. The service is REST adherent and can be used with any programming language.

Problem Statement

Direct marketing, either through mail, email, phone, etc., is a common tactic to acquire customers. Because resources and a customer's attention is limited, the goal is to only target the subset of prospects who are likely to engage with a specific offer. Predicting those potential customers based on readily available information like demographics, past interactions, and environmental factors is a common machine learning problem. The essence of the API is to predict if a customer will click on an Ad or not after pushing out campaigns.

Setup

Set up the flask application environment. In the same environment, there is a Preprocessing.py and Postprocess.py scripts. The Preprocessing.py does the data cleaning and preprocessing. The Postprocess.py loads the preprocessor pikle object which transforms the data. The model folder contains a model.h5 and the preprocessor.pkl object. The Features.yml contains features used for building the model.

Flask
        |__ App.py
        |__ Features.yml
        |__ Preprocessing.py
        |__ Postprocess.py
        |__ Model
          |__model.h5
          |__ preprocessor.pkl

Pre/Post-Processing

Flask TensorFlow Serving supports the following Content-Types for requests: *application/json (default) *text/csv

The flask app will convert data in these formats to TensorFlow Serving REST API requests, and will send these requests to the default serving signature of our Saved Model. Read in the yaml file but you need to specify the path to your yaml file.

Expected Output

The expected output of this application are probability scores of users, these scores ranks their individual propensities to click on a digital campain Advertisement. This users are ranked on a scale of 0 - 1. Users with high probabilities have high propensity to click on the campaign adverts while users with low probabilities have low propensity.