DA5030 Practice 2 | Machine Learning & Data Mining Solved

1.  The built-in dataset USArrests contains statistics about violent crime rates in the US States. Determine which states are outliers in terms of murders. Outliers, for the sake of this question, are de         ned as values that are more than 1.5 standard deviations from the mean.

2.  For the same dataset as in (1), is there a correlation between urban population and murder, i.e., as one goes up, does the other statistic as well? Comment on the strength of the correlation. Calculate the Pearson coe         cient of correlation in R.

3.  Based on the data on the growth of mobile phone use in Brazil (you'll need to copy the data and create a CSV that you can load into R or use the gsheet2tbl() function from the gsheet package), forecast phone use for the next time period using a 2-year weighted moving average (with weights of

5 for the most recent year, and 2 for other), exponential smoothing (alpha of 0.4), and linear regression trendline.

4.  Calculate the squared error for each model, i.e., use the model to calculate a forecast for each given time period and then the squared error. Finally, calculate the average (mean) squared error for each model. Which model has the smallest mean squared error (MSE)?

5. Calculate a weighted average forecast by averaging out the three forecasts calculated in (3) with the following weights: 4 for trend line, 2 for exponential smoothing, 1 for weighted moving average. Remember to divide by the sum of the weights in a weighted average