Abstract

• This particular work aims to analyze a very importantsubject - adoptability of pets. The paper uses PetFinder.myAdoption Prediction dataset on Kaggle for analysis andtries to incorporate different modalities to predict speedof adoption. Extensive data analysis is used to figure outfavourable traits in pets since knowing beforehand thecharacteristics of adoptable dogs may help shape shelterpolicy. The paper also discusses different feature extractiontechniques for different modalities and compares variousmodels to obtain the best results.

Requirements

• python3, sklearn, keras, pickle, numpy, nltk, scikit, lime, openCV, pandas, LGBM.

Data

• all the data and features can be found in the drive https://drive.google.com/drive/folders/18MiUede2DyPXPXd9NwYbfDIXreVEWuTL?usp=sharing

Jupyter notebooks

• contains all the jupyter notebooks used
• Dataset_description.ipynb contains the dataset description
• Text_Analysis_Model.ipynb contains machine learning models and text analysis
• try_lime.ipnb contains lime visualization of features.
• train_test.ipynb contains keras models containing all the fusion techniques.

Code

• Contains all preprocessing scripts used.

Project Proposal and Paper

• contains all the papers followed and project proposal made

Data_Analysis.html

• gives a simple, compact and good visualiszation of whola dataset.

Thank you. Team Members : Rahul Kukreja, Anunay Yadav, Ansh Kumar Sharma