Slide 7
Slide 7 text
Test Set:
The test set here will be the actual fixtures of the tournament in the group stage, which looks like
this. So there are two cases here, assuming each team has the 50% possibility of batting or fielding first.
The green column indicates target variable and the red columns indicate the missing values.
Case 1: Team A bats first
Match
Inning
s Team A Team B RR Score Wickets
Over
s ground venue result
1 1 Australia England Melbourne home ?
2 1
New
Zealand Sri Lanka Christchurch
home ?
. . . . . . . . . ?
To fill the missing values Exponential Moving Averages(EMA) is used. For Example,
(, ) =
(
, 3) + (
, 5) + (
, 4)
3
Exponential average provides more weightage to recent performances. In the similar way other column
values were calculated.
The exponential moving average is calculated using MATLAB.
Score=zeros(length(TeamA),1)
for n=2:length(TeamA)
teamA_teamB_rows = strcmp(cricket.TeamA,TeamA{n}) &
strcmp(cricket.TeamB,TeamB{n});
teamA_rows = strcmp(cricket.TeamA,TeamA{n});
recent = cricket.Score(teamA_teamB_rows);
complete = full_data.Score(teamA_teamB_rows);
overall = cricket.Score(teamA_rows);
if length(complete)>6
x = tsmovavg(recent,'e',3,1);
y = tsmovavg(complete,'e',5,1);
z = tsmovavg(overall,'e',5,1);
exp = (x(end)+y(end)+z(end))/3 ;
else
exp = (mean(complete) + mean(overall))/2;
end
Score(n,1) = exp(end);
fprintf('%f\n',exp(end));
end