Airbnb-Data-Prediction

Airbnb's program to predict the country of reservation for new subscribers

Overview

📕Data: https://www.kaggle.com/c/airbnb-recruiting-new-user-bookings

🔑Purpose: Predicting Reservation Destinations. Building a Machine Learning Model

👩🏻‍💻 Datasets Used: train_users_2.csv, test_users.csv

Data Preprocessing

Describe the preprocessing steps taken for the numerical(age, sighnup_flow) and categorical data(gender, language). (Fills a null value and removes unfilled rows) If the age is less than 10 years old or more than 90 years old, it was considered an abnormal value and dropped
Final Preprocessing Method:

Scaling

the numerical data was preprocessed by scaling (MinMax, Robust, Standard)

Encoding

the categorical data was preprocessed by encoding (OneHot, Ordinal, Label).

Data Visualization

Visualization based on data frames processed for outliers

Modeling

Describe the models used (Logistic Regression and Decision Tree)

Model Visualization

Evaluation

Use k fold cross validation to find the optimal k value

Use confusion matrices to visualize the results of classification tasks and evaluate the performance of models

Logistic Regression decision tree

Best 5 combination

Extract the top 5 combinations of all combinations considering 3 encoders (One-hot, Ordinary, Label) + 3 Scaling (MinMax, Standard, Robust) + hyperparameters

Analysis

Analyze critical features in predicting destinations through modeling

Usage

It can be used to interpret other data because it has carried out an end-to-end process that includes all the steps, including data collection, preprocessing, modeling, evaluation, and data analysis results analysis.