paudiaz Goto Github PK

assessing-fit-polynomial-regression-

Assessing Fit (polynomial regression)

clustering

I use here nearest neighbors and clustering to retrieve documents that interest users, by analyzing their text. I explored two document representations: word counts and TF-IDF. This notebook retrieve articles from Wikipedia about famous people.

cross-validation-

In this mini-project, I made a training-testing split in the data. This is the first step toward the final project, of building a POI identifier by using decision tree classifier taking care of its accuracy.

data-science-intensive-course.-capstone-project

This project deals with the always painful process of moving to a new city. A recommender model based on an unsupervised approach is here presented for the city of Washington DC (US). The work includes the data wrangling process, a statistical analysis of the criminal activity recorded along 2011 to 2013 and finally the use of machine learning to build an unsupervised recommender system based on the preferences of the user and on four main features: safety, access to public transportation services, education and economic indexes.

data_wrangling_json

data_wrangling_xml

datastoryi

datastoryii

deep-learning-image-retrieval

We build here a new image retrieval models and explore their results on different parts of our image dataset.

dimensionality-reduction-pca

eigen faces

examining-racial-discrimination-in-the-us-job-market

In this exercise, You will perform a statistical analysis to establish whether race has a significant impact on the rate of callbacks for resumes.Instructions and data are provided in the attached ipython notebook.

feature_selection

intensivedatascience

All the scripts here posted have been coded during the SlideRules workshop entitled "Intensive Data Science".

introduction-to-decision-tree-algorithms

This is the code to accompany the Lesson 3 (decision tree) mini-project.

k_means

The classic k-means clustering algorithm

knn-regression

lasso-regresion

Lasso Regression

linear_classifier

Exploring logistic regression and feature engineering with existing GraphLab Create functions.

logisticregression

Logistic Regression from scratch

metrics-evaluation

Here we use recall_score, precision_score and f1_score to evaluate how good is our identifier algorithm.

multiple-linear-regression

Multiple Regression (gradient descent)

naive-bayes-e-mail-identifier

We have a set of emails, half of which were written by one person and the other half by another person at the same company . Our objective is to classify the emails as written by one person or the other based only on the text of the email. We will use Naive Bayes. We are also interested in checking the training and predicting time that this algorithm takes.

paying-debt-off-in-a-year

This program calculates the minimum fixed monthly payment needed in order pay off a credit card balance within 12 months. By a fixed monthly payment, we mean a single number which does not change each month, but instead is a constant amount that will be paid each month. The following variables contain values as described below: balance - the outstanding balance on the credit card annualInterestRate - annual interest rate as a decimal The program prints out one line: the lowest monthly payment that will pay off all debt in under 1 year, for example: Lowest Payment: 180 Assume that the interest is compounded monthly according to the balance at the end of the month (after the payment for that month is made). The monthly payment must be a multiple of $10 and is the same for all months. Notice that it is possible for the balance to become negative using this payment scheme, which is okay. A summary of the required math is found below: Monthly interest rate = (Annual interest rate) / 12.0 Monthly unpaid balance = (Previous balance) - (Minimum fixed monthly payment) Updated balance each month = (Monthly unpaid balance) + (Monthly interest rate x Monthly unpaid balance)

regression

This script deals with regression. The goal here is to perform a house price estimation using regression models with one and multiple features. To train the model I used training/test set splitting. Finally the root mean squared errors (RMSE) of each example are compared to check which method predicts a more accurate price for some target houses.

regularization

ridge-regresion

Ridge Regression (gradient descent)

sentiment_analysis

Here I deploy some models that predict a class (positive/negative sentiment) from input features (text of the reviews, user profile information,...).This task is an example of classification. I also analyzed the accuracy of the classifiers using roc_curve metrics.

simple-linear-regression

Prediction of house prices using simple (one input) linear regression

song_recommender_system

Here a building recommender system is build to find products, music that interest users. I compared the simple popularity-based recommendation with a personalized model, and showed the significant improvement provided by personalization.

statistics-project-3