Lesson 8 - Data Science

What are the 3 main data science problems?

• Regression

• Clustering

• Classification

What is regression?

• Finds a relationship between different data points

• Trend between 2 vars measured in same environment

• Ex: finding rate of spread of diseases (new cases vs. time)

What is clustering?

• Method of grouping closely related data

• Used to find patterns and anomalies in data

• Ex: group customers based on purchase behavior

What is classification?

• Sorting of data into specific categories or groups

• Train machine learning model by labeling data into desired categories

• Ex: sort products as popular or unpopular

Precision and Recall Example

What classification algorithm do we use and what are advantages/disadvantages?

K Nearest Neighbors (KNN)

• finds desired number of closest neighbors and assigns the classification corresponding to the most common one returned

• advantages: no training period, easy implementation

• disadvantages: difficult to scale for large data sets, sensitive to missing and noisy data

What clustering algorithm do we use and what does it do?

K-Means

• divides data into clusters by minimizing the distance between each point and the cluster’s centroid

• after all assignments done it recalculates the centroid and repeats the process until it finds true centroid