Pankush Kukreja's repositories
Stock-Prediction-using-LSTM
I will be considering the google stocks data and will create a LSTM network for prediction.
Image-Classification-using-CNN
This is the Kaggle dataset for Image classification of Dog and Cat. i have considered 5000 images out of 25000 image. Data is manually been divided as 4000 image of each class in training and 1000 image as testing.
Weight-of-Evidence-and-IV-Values
Implementation fo WO and IV Values in python
Time-Series---Jetrails-problems
Using this jet rails real life problem, i tried explaining all the time series algorithms to get better understanding of time series.
Annual-Income-Prediction
This is a prediction model to predict annual income for an adult whether it will be more or less than 50K. and explaining all the classification model evaluation metrics and curve like ROC, AUC, Precision _recall, Gini, Ks_score, Logloss, Concordance.
Basic-ANN
Here i am using Bank data which will predict if he customer will leave Bank or not using Artificial Neural network.
Book-Recommendation-system
Book recommendation collaborative filtering system using Pearson correlation and nearest neighbor algorithm
CatBoost-Implementation
Implementation of Cat Boost using Annual Income prediction data.
Credit-Card-Deault-Assesment
This is data for customers using Credit Card , having 25 columns with their Eductaion, Martial Status, Age , Pays and previous Bill and payments for cc, so based on this information we will predict whether in future the customer will default the CC payment or not. I am using XGBoost for model, Randomised search for parameter tuning and also made it in 3 hidden layer Artificial neural network with SGD optimizer.
Cross-Validation-and-Its-Type
This is the implementation of all type of Cross validation and Back ward elimination technique
EDA-using-Wine-Quality-data-set
In this i explained all aspects of Exploratory data analysis using Wine Quality data set from kaggle
House-Prices-Advanced-Regression-Techniques
This is Kaggle challenge for house prediction, it has alot of missing values which needs to be cleaned. i have used Regression ,Boosting, and Deep Neural network and tuning them.
Human-Activity-Classification
The project involves training a Machine Learning model to classify the kind of activity a person is performing including sitting, standing, laying, walking, walking upstairs and walking downstairs using data collected from smartphones.
kaggle-Credit-Card-Fraud-Detection
This is an highly imbalanced data with only 1.72% minority and 98.28% majority class, i will be explaining Up and down sampling and effect of sampling before and while doing cross validation. Model has been evaluated using precision recall curve.
Natural-Language-Processing-NLP-
In this you can see the implementation of basics preprocessing of text from paragraph, Bag of words and TF-IDF.
Pandas-Tutorials
This is how i expertise in Pandas
PipelineRepo
Pipeline purpose
recommendation
Recommendation System using ML and DL
Recommendation-System
Explaining recommendation and its type
Restro_Review_Analysis-using-NLP
I have 1000 reviews for a restaurant based on that i will classify them as 1(Good) and 1(bad).
Sales-Prediction
Monthly sales prediction of a furniture store using past daily order data.
SMS-SPam-Classification
i will be processing 5000 sms and will classify them as Spam or not
Task-Manager-App-Using-Flask
A simple flask app demostrating the Crud functions in flask.
Twitter_senitiment_analysis-using-word2vec-and-ann
This is solution for Analytics Vidhya Data hackthon on Twitter_Senitiment_Analysis.
Web_Scraping_BS4
Web scraping using python Beautiful Soup and Request library
XGB-vs-LGBM-vs-CATBoost
Using kaggle flight delay data to explain the difference between XGB, LGBM and Catboost