Variable names - Githubissues

Below are the variable names I am using. Feel free to correct any typo or change the name. You may comment your changes.

General Settings

{
    "study_name": "HeartSteps",
    "study_description": "Practice with HeartSteps using Scenario 1 where you are using published results from an MRT",
    "personalized_scenario": "Scenario 1: I have published results from MRT",
    "proximal_outcome_name": "(log) Next 30 min step count",
    "personalization_method": "Thompson Sampling - Two intervention option",
    "intervention_component_name": "Activity suggestion"
}

Intervention Settings

{
    "update_day": "Daily",
    "condition_1": "Currently walking or running",
    "condition_2": "Finished an activity in the prev 90 min",
    "update_hour": "12:00am",
    "intervention_option_a": "Activity Suggestion",
    "intervention_option_b": "Do Nothing: No activity suggestion",
    "decision_point_frequency": "5",
    "decision_point_frequency_time": "Day",
    "intervention_probability_lower_bound": "0.1",
    "intervention_probability_upper_bound": "0.8"
}

Model Settings

{
    "intercept_prior_mean": "0",
    "max_proximal_outcome": "8.0",
    "min_proximal_outcome": "-0.69",
    "treatment_prior_mean": "0.13",
    "proximal_outcome_name": "(log) Next 30 min step count",
    "proximal_outcome_type": "Continuous",
    "intervention_component_name": "Activity suggestion",
    "intercept_prior_standard_deviation": "10",
    "treatment_prior_standard_deviation": "0.07"
}

Covariates

{
    "2c201b1d-da6c-40e0-9b61-1da2ba81dfb1": {
        "covariate_name": "(log) Prior 30 min step count",
        "covariate_type": "Continuous",
        "covariate_max_val": "8.0",
        "covariate_min_val": "-0.69",
        "tailoring_variable": "no",
        "proximal_outcome_name": "(log) Next 30 min step count",
        "main_effect_prior_mean": "0",
        "intervention_component_name": "Activity suggestion",
        "main_effect_prior_standard_deviation": "10"
    },
    "e96b404f-91c1-4f80-9228-96c6b4b2f594": {
        "covariate_name": "Location",
        "covariate_type": "Binary",
        "covariate_max_val": "1",
        "covariate_min_val": "0",
        "tailoring_variable": "yes",
        "covariate_meaning_0": "Home/Work",
        "covariate_meaning_1": "Unknown location",
        "proximal_outcome_name": "(log) Next 30 min step count",
        "main_effect_prior_mean": "01",
        "intervention_component_name": "Activity suggestion",
        "interaction_coefficient_prior_mean": "0",
        "main_effect_prior_standard_deviation": "101",
        "interaction_coefficient_prior_standard_deviation": "10"
    }
}

General Settings { "proximal_outcome_name": "(log) Next 30 min step count", => The name of the "proximal_outcome" "personalization_method": "Thompson Sampling - Two intervention option", => "ThompsonSampling" "intervention_component_name": "Activity suggestion" => The name of the "decision" } Intervention Settings { "condition_1": "Currently walking or running", => name of the "eligibility" "condition_2": "Finished an activity in the prev 90 min", => name of the "eligibility" "update_day": "Daily", => when function "update" is called "update_hour": "12:00am", => when function "update" is called "intervention_option_a": "Activity Suggestion", => name of "decision_options" with index 1 "intervention_option_b": "Do Nothing: No activity suggestion", => name of "decision_options" with index 0 "intervention_probability_lower_bound": "0.1", => "lower_clip" "intervention_probability_upper_bound": "0.8" => "upper_clip" } Model Settings {

"max_proximal_outcome": "8.0", => This is the upper bound for "proximal_outcome" (We can name this "proximal_outcome_max")
"min_proximal_outcome": "-0.69", => This is the lower bound for "proximal_outcome" (We can name this "proximal_outcome_min"
"proximal_outcome_type": "Continuous", => This is to check if the "proximal_outcome" is float
"intercept_prior_mean": "0", => "alpha_0_mu_bias"
"intercept_prior_standard_deviation": "10", => "alpha_0_sigma_bias"
"treatment_prior_mean": "0.13", => "beta_mu_bias"
"treatment_prior_standard_deviation": "0.07" => "beta_sigma_bias"

} Covariates { "2c201b1d-da6c-40e0-9b61-1da2ba81dfb1": { "covariate_name": "(log) Prior 30 min step count", => "feature_name" "covariate_type": "Continuous", => this is to check if "feature" is float "covariate_max_val": "8.0", => This is the upper bound for this feature (We can name this "feature_max") "covariate_min_val": "-0.69", => This is the lower bound for this feature (We can name this "feature_min") "tailoring_variable": "no", => "beta_selected_features" "main_effect_prior_mean": "0", => "alpha0_mu" "main_effect_prior_standard_deviation": "10" => "alpha0_sigma" }, "e96b404f-91c1-4f80-9228-96c6b4b2f594": { "covariate_name": "Location", => "feature_name" "covariate_type": "Binary", => this is to check if "feature" is binary "covariate_max_val": "1", => This is the upper bound for this feature (We can name this "feature_max") "covariate_min_val": "0", => This is the lower bound for this feature (We can name this "feature_min") "tailoring_variable": "yes", => "beta_selected_features" "covariate_meaning_0": "Home/Work", "covariate_meaning_1": "Unknown location", "main_effect_prior_mean": "01", => "alpha0_mu" "main_effect_prior_standard_deviation": "101", => "alpha0_sigma" "interaction_coefficient_prior_mean": "0", => "beta_mu" "interaction_coefficient_prior_standard_deviation": "10" => "beta_sigma" } }

Notes:

I will need the min and max for the "proximal_outcome" and all the "features" in the algorithm. I have not implemented it yet and you may not find these in the algorithm, but it's one of the todo's.
Notice that if a covariate is not a tailoring variable, it will be missing "beta_mu" and "beta_sigma." Please set those as "nan."
In the algorithm, there are also "noise_scale" and "noise_degree." I apologized that I only realized it now that this is something I need to mention to you guys. Recall that a while ago I said there will be two scenarios -- "using MRT paper" versus "using MRT data." Ali is currently implementing "using MRT paper." In this scenario, we won't be asking for these two parameters and we will define the two parameters for people. However, in the scenario of "using MRT data," we will be asking for these two parameters. I imagine we will have one more page right after "Covariate" in "Model Settings" on the user interface. Please let me know what's the best way to help you with this. I can have a page holder for this, and please hide it in the scenario of "using MRT paper."

Please let me know if anything is unclear. Thanks so much!

mDOT-Center / pJITAI

Variable names #46