compare_results: Identifies method with least error

In RashiMohta/COVID-19-cases-prediction: Jump-drop adjusted prediction of cumulative infected cases using the modified SIS model

Identifies method with least error

Description

Compares the validation rmse using original and adjusted data and returns the one with lesser error.

Usage

compare_results(
  population = 18710922,
  gamma = 1/14,
  cur_date = "2020-5-29",
  start_date = "2020-3-13",
  last_n_day = 20,
  last_limit = 30,
  next_n_days = 20,
  data,
  ub_for_adjustment = 5,
  bound_metric = "C3_1day",
  df_confirmed_values,
  mu
)

Arguments

population	number of people in the state
gamma	recovery rate
cur_date	current date for start of prediction phase
start_date	start date in the considered dataset
last_n_day	number of days in training phase
last_limit	maximum number of days in the validation period
next_n_days	number of days in the prediction phase
data	state wise daily cases adjusted data for the given state
ub_for_adjustment	upper bound for the duration of a jump or drop
bound_metric	C1, C2, or C3 metric can be selected
df_confirmed_values	Dataframe of dates and observed number of daily confirmed cases
mu	mortality rate of the infection

Value

Returns graph showing observed and predicted cummulative cases using the best method and the corresponding MSE for validation and prediction period.

Note

This function is called by the user. It prints the MSE values of predictions made with and without making the adjustments and returns the best predictions.

RashiMohta/COVID-19-cases-prediction documentation built on Oct. 26, 2024, 9:48 a.m.