$8\mathrm{.}9$ LAB: Logistic regression using logit$()$

Using the csv file nbaallelo$\_$log$.$csv and the logit function, construct a logistic regression model to classify whether a team will win or lose a

game based on the team's elo$\_$i score.

Read in the file nbaaello$\_$log$.$csv$.$

The target feature will be converted from string to a binary feature by the provided code.

Split the data into $70$ percent training set and $30$ percent testing set. Set random$\_$state $=0.$

Use the logit function to construct a logistic regression model with wins as the target and elo$\_$i as the predictor.

Print the coefficients of the model.

Ex: If the feature pts is used as the predictor, rather than elo$\_$i$,$ the output is:

Optimization terminated successfully.

Current function value: $0\mathrm{.}621201$

Iterations $5$

Intercept $-5\mathrm{.}908580$

pts $0\mathrm{.}057528$

dtype: float$64$

main.pyload nbaallelo$\_$log$.$csv into a dataframe

df $=$ # code to load csv filewins $=$ df$.$game$\_$result $=="$W$"$

bool$\_$val $=$ np$.$multiply$($wins$,1)$

wins $=$ pd$.$DataFrame$($bool$\_$val, columns $=["$game$\_$result"$])$

wins$\_$new $=$ wins.rename$($columns $=\{"$game$\_$result": "wins"$\})$

df$\_$final $=$ pd$.$concat$([$df$,$ wins$\_$new$],$ axis$=1)$train$,$ test $=$ # code to split df$\_$final into training and test setslm$\}=\backslash$mathrm$\{$ # code to construct logistic model using the logit functionprint$($# code to return coefficients$)$