I want to train the AWS xgBoost model for classification into two categories based on my dataset which contains 1000 Emails and the value to predict. Ex data - abc@gmail -- high 123@yahoo -- low xyz@gmail -- low jkl@yahoo -- high For training the model on 1000 rows of such data, as xgboost only expects numerical data, I had to use pd.getDummies() which converts this into 1000 columns by 1000 rows of data. Converts the result column into two - isHigh, isLow with values 0 and 1. Now when I want predictions from my model, it expects an input of 1000 parameters for the REST API(which I created for predictions), ideally, I would like to provide an email as the parameter and get a result of high/low or isHigh value between 0 and 1. Is the choice of Algorithm wrong? Is there any other way so that when I ask the model to provide predictions it does not want 100 parameters? Continue reading...
