poisson_simulate_matches: Poisson Simulate Matches

In neilcuz/panenkar: Panenkar

Description

Predict multiple matches using Poisson regression, updating the training set in the process

Usage

poisson_simulate_matches(
  training_set,
  test_set,
  xi = 0.0016,
  max_goals = 8,
  markets = c("result", "over_under", "both_teams_to_score"),
  over_under_goals = 2.5,
  zero_inflated = FALSE,
  weight_cut_off = NA
)

Arguments

training_set,	columns of which must be "home_team", "away_team", "home_goals", "away_goals", "match_date"
test_set,	same column restrictions as training_set
xi	parameter for the exponential time weighting, by default 0.0016
max_goals,	maximum number of home goals and/or away goals to model
markets,	vector contain one or more of "result", "over_under" or "both_teams_to_score"
over_under_goals	typically n.5 where n in 1, 2, 3 .. etc, default = 2.5. Argument only used if "over_under" listed in markets
zero_inflated	default FALSE, if TRUE the zero inflated poisson model is used
weight_cut_off,	NA by default can set a lower cut off e.g. 0.01 which will drop observations from model data set to potentially improve speed with a potential accuracy trade off

Value

probabilities for outcomes of the various markets for each match in the test set

neilcuz/panenkar documentation built on June 19, 2021, 7:31 p.m.