$1.$ Define unsupervised versus supervised data mining

Identify which type of data mining you would use to run the following four projects

$2.$ Autonomous car driving dataset with crash occurrences identified in the data, to determine what are the most critical factors to monior to avoid a crash.

$3.$ Population and employment data for Ohio, with a desire to understand where people like to work and live in the state.

$4.$ Monthly and daily records of product sales across various states, cities and stores with a desire to understand why some stores are selling more products than others.

$5.$ Tire failure dataset on a certain brand of tires in the US with car operational data, tire failurebrecords and weather data, with a desire to understand the causes for failure.

Model training and testing datasets

$6.$ Should we strive for the highest possible accuracy with the training set? Why or why not?

$7.$ Explain why we sometimes need to balance model training data.

For the next two questions and for modeling training purposes, we are running a fraud classification model and have initially identified a training data set of $10,000$ records, of which we know $400,$ or $4\%,$ are fraudulent.

$8.$ To better balance the training dataset, how many fraudulent records need to be added if we want the proportion of fraudulent records to be $20\%?$

$9.$ How many total records would then be in our training dataset?

Final Question

$10.$ When should a test data set be balanced?