{"main_mean":{"name":"Predicted number of fatalities","description":"Point prediction of the number of fatalities in impending conflict, presented per type of violence and level of analysis.","unit":{"name":"Fatalities","range":"0 - unlimited"},"original":"sc_cm_{tv}_main","type":"Ensemble model","visualization":{"legend":"Predicted fatalities","colormap":"rainbow","dictionary":{"cm":"dictionary_stand_1p_10k","pgm":"dictionary_stand_1p"},"transformation":{"data":"ln","popup":"N/A","suffix":""}}},"main_mean_ln":{"name":"Predicted log number of fatalities","description":"Point prediction of the log number of fatalities in impending conflict, presented per type of violence and level of analysis. Expressed in natural logged form plus 1 [ln(Y+1)].","unit":{"name":"ln(fatalities)","range":"0 - unlimited"},"original":"sc_cm_{tv}_main_ln","type":"Ensemble model","visualization":{"legend":"Predicted fatalities","colormap":"rainbow","dictionary":{"cm":"dictionary_stand_1p_10k","pgm":"dictionary_stand_1p"},"transformation":{"data":"N/A","popup":"N/A","suffix":""}}},"main_dich":{"name":"Predicted probability of conflict","description":"Predicted probability of at least 25 battle-related deaths (BRDs) per country-month, or at least 1 BRD per PRIO-GRID-month. Estimated from the ensemble model results for the applicable type of violence and level of analysis.","unit":{"name":"Probability","range":"0-1"},"original":"sc_cm_{tv}_dich_main","type":"Non-applicable","visualization":{"legend":"Predicted probability","colormap":"rainbow","dictionary":{"cm":"dictionary_dichotomous_cm","pgm":"dictionary_dichotomous_pgm"},"transformation":{"data":"logit","popup":"*100","suffix":"%"}}},"surrogate_mean_ch_ln":{"name":"Predicted log fatalities, isolated contribution from conflict history features","description":"Predicted log fatalities per type of violence (tv) and level of analysis, given only the history of the respective type of violence (as recorded by the UCDP GED and UCDP Candidate datasets) in the country or PRIO-GRID-cell at hand. Expressed in natural logged form plus 1 [ln(Y+1)].

The prediction is generated by a surrogate or interpretation model that shows how much of the main VIEWS conflict prediction can be related to the input variable `ln_ged_{tv}_dep`; in practice, a regression model with the main prediction as the dependent variable and a flexible function of `ln_ged_{tv}_dep` as the independent one.

The surrogate model results should be seen as an indication of the relative importance of specific independent variables (predictors), rather than point predictions of their own (they do not sum up to the point predictions from the main ensemble model).

Learn more about the surrogate models.","unit":{"name":"ln(fatalities)","range":"0 - unlimited"},"original":"sc_cm_{tv}_surrogate_ch_ln","type":"Surrogate model","visualization":{"legend":"Predicted fatalities","colormap":"rainbow","dictionary":{"cm":"dictionary_stand_1p_10k","pgm":"dictionary_stand_1p"},"transformation":{"data":"N/A","popup":"N/A","suffix":""}}},"surrogate_mean_ch":{"name":"Predicted fatalities, isolated contribution from conflict history features","description":"Predicted fatalities per type of violence (tv) and level of analysis, given only the history of the respective type of violence (as recorded by the UCDP GED and UCDP Candidate datasets) in the country or PRIO-GRID-cell at hand.

The prediction is generated by a surrogate or interpretation model that shows how much of the main VIEWS conflict prediction can be related to the input variable `ln_ged_{tv}_dep`; in practice, a regression model with the main prediction as the dependent variable and a flexible function of `ln_ged_{tv}_dep` as the independent one.

The surrogate model results should be seen as an indication of the relative importance of specific independent variables (predictors), rather than point predictions of their own (they do not sum up to the point predictions from the main ensemble model).

Learn more about the surrogate models.","unit":{"name":"Fatalities","range":"0 - unlimited"},"original":"sc_cm_{tv}_surrogate_ch","type":"Surrogate model","visualization":{"legend":"Predicted fatalities","colormap":"rainbow","dictionary":{"cm":"dictionary_stand_1p_10k","pgm":"dictionary_stand_1p"},"transformation":{"data":"ln","popup":"N/A","suffix":""}}},"surrogate_mean_dem_ln":{"name":"Predicted log fatalities, isolated contribution from the liberal democracy index","description":"Predicted log fatalities per level of analysis, given only the liberal democracy index in the country or PRIO-GRID-cell at hand. Expressed in natural logged form plus 1 [ln(Y+1)].

The prediction is generated by a surrogate or interpretation model that shows how much of the main VIEWS conflict prediction can be related to the input variable `vdem_v2x_libdem`; in practice, a regression model with the main prediction as the dependent variable and a flexible function of `vdem_v2x_libdem` as the independent one.

The surrogate model results should be seen as an indication of the relative importance of specific independent variables (predictors), rather than point predictions of their own (they do not sum up to the point predictions from the main ensemble model).

Learn more about the surrogate models.","unit":{"name":"ln(fatalities)","range":"0 - unlimited"},"original":"sc_cm_{tv}_surrogate_dem_ln","type":"Surrogate model","visualization":{"legend":"Predicted fatalities","colormap":"rainbow","dictionary":{"cm":"dictionary_stand_1p_10k","pgm":"dictionary_stand_1p"},"transformation":{"data":"N/A","popup":"N/A","suffix":""}}},"surrogate_mean_dem":{"name":"Predicted fatalities, isolated contribution from the liberal democracy index","description":"Predicted fatalities per level of analysis, given only the liberal democracy index in the country or PRIO-GRID-cell at hand.

The prediction is generated by a surrogate or interpretation model that shows how much of the main VIEWS conflict prediction can be related to the input variable `vdem_v2x_libdem`; in practice, a regression model with the main prediction as the dependent variable and a flexible function of `vdem_v2x_libdem` as the independent one.

The surrogate model results should be seen as an indication of the relative importance of specific independent variables (predictors), rather than point predictions of their own (they do not sum up to the point predictions from the main ensemble model).

Learn more about the surrogate models.","unit":{"name":"Fatalities","range":"0 - unlimited"},"original":"sc_cm_{tv}_surrogate_dem","type":"Surrogate model","visualization":{"legend":"Predicted fatalities","colormap":"rainbow","dictionary":{"cm":"dictionary_stand_1p_10k","pgm":"dictionary_stand_1p"},"transformation":{"data":"ln","popup":"N/A","suffix":""}}},"surrogate_mean_pop_ln":{"name":"Predicted log fatalities, isolated contribution from population size feature","description":"Predicted log fatalities per type of violence (tv) and level of analysis, given only the population size in the country at hand. Expressed in natural logged form plus 1 [ln(Y+1)].

The prediction is generated by a surrogate or interpretation model that shows how much of the main VIEWS conflict prediction that can be related to variable `wdi_sp_pop_totl`; in practice, a regression model with the main prediction as the dependent variable and a flexible function of `wdi_sp_pop_totl` as the independent one.

The surrogate model results should be seen as an indication of the relative importance of specific independent variables (predictors), rather than point predictions of their own (they do not sum up to the point predictions from the main ensemble model).

Learn more about the surrogate models.","unit":{"name":"ln(fatalities)","range":"0 - unlimited"},"original":"sc_cm_{tv}_surrogate_pop_ln","type":"Surrogate model","visualization":{"legend":"Predicted fatalities","colormap":"rainbow","dictionary":{"cm":"dictionary_stand_1p_10k","pgm":"dictionary_stand_1p"},"transformation":{"data":"N/A","popup":"N/A","suffix":""}}},"surrogate_mean_pop":{"name":"Predicted fatalities, isolated contribution from population size feature","description":"Predicted fatalities per type of violence (tv) and level of analysis, given only the population size in the country at hand.

The prediction is generated by a surrogate or interpretation model that shows how much of the main VIEWS conflict prediction that can be related to variable `wdi_sp_pop_totl`; in practice, a regression model with the main prediction as the dependent variable and a flexible function of `wdi_sp_pop_totl` as the independent one.

The surrogate model results should be seen as an indication of the relative importance of specific independent variables (predictors), rather than point predictions of their own (they do not sum up to the point predictions from the main ensemble model).

Learn more about the surrogate models.","unit":{"name":"Fatalities","range":"0 - unlimited"},"original":"sc_cm_{tv}_surrogate_pop","type":"Surrogate model","visualization":{"legend":"Predicted fatalities","colormap":"rainbow","dictionary":{"cm":"dictionary_stand_1p_10k","pgm":"dictionary_stand_1p"},"transformation":{"data":"ln","popup":"N/A","suffix":""}}},"surrogate_mean_imr_ln":{"name":"Predicted log fatalities, isolated contribution from infant mortality rate","description":"Predicted log fatalities per type of violence (tv) and level of analysis, given only the infant mortality rate in the country at hand. Expressed in natural logged form plus 1 [ln(Y+1)].

The prediction is generated by a surrogate or interpretation model that shows how much of the main VIEWS conflict prediction that can be related to variable `wdi_sp_dyn_imrt_in`; in practice, a regression model with the main prediction as the dependent variable and a flexible function of `wdi_sp_dyn_imrt_in` as the independent one.

The surrogate model results should be seen as an indication of the relative importance of specific independent variables (predictors), rather than point predictions of their own (they do not sum up to the point predictions from the main ensemble model).

Learn more about the surrogate models.","unit":{"name":"ln(fatalities)","range":"0 - unlimited"},"original":"sc_cm_{tv}_surrogate_imr_ln","type":"Surrogate model","visualization":{"legend":"Predicted fatalities","colormap":"rainbow","dictionary":{"cm":"dictionary_stand_1p_10k","pgm":"dictionary_stand_1p"},"transformation":{"data":"N/A","popup":"N/A","suffix":""}}},"surrogate_mean_imr":{"name":"Predicted fatalities, isolated contribution from infant mortality rate","description":"Predicted fatalities per type of violence (tv) and level of analysis, given only the infant mortality rate in the country at hand.

The prediction is generated by a surrogate or interpretation model that shows how much of the main VIEWS conflict prediction that can be related to variable `wdi_sp_dyn_imrt_in`; in practice, a regression model with the main prediction as the dependent variable and a flexible function of `wdi_sp_dyn_imrt_in` as the independent one.

The surrogate model results should be seen as an indication of the relative importance of specific independent variables (predictors), rather than point predictions of their own (they do not sum up to the point predictions from the main ensemble model). Learn more about the surrogate models.","unit":{"name":"Fatalities","range":"0 - unlimited"},"original":"sc_cm_{tv}_surrogate_imr","type":"Surrogate model","visualization":{"legend":"Predicted fatalities","colormap":"rainbow","dictionary":{"cm":"dictionary_stand_1p_10k","pgm":"dictionary_stand_1p"},"transformation":{"data":"ln","popup":"N/A","suffix":""}}},"surrogate_mean_nch_ln":{"name":"Predicted log fatalities, isolated contribution from neighborhood conflict history","description":"Predicted log fatalities per type of violence (tv) and level of analysis, given only the neighborhood conflict history in the country at hand. Expressed in natural logged form plus 1 [ln(Y+1)].

The prediction is generated by a surrogate or interpretation model that shows how much of the main VIEWS conflict prediction that can be related to variable `splag_1_decay_ged_{tv}_5`; in practice, a regression model with the main prediction as the dependent variable and a flexible function of `splag_1_decay_ged_{tv}_5` as the independent one.

The surrogate model results should be seen as an indication of the relative importance of specific independent variables (predictors), rather than point predictions of their own (they do not sum up to the point predictions from the main ensemble model).

Learn more about the surrogate models.","unit":{"name":"ln(fatalities)","range":"0 - unlimited"},"original":"sc_cm_{tv}_surrogate_nch_ln","type":"Surrogate model","visualization":{"legend":"Predicted fatalities","colormap":"rainbow","dictionary":{"cm":"dictionary_stand_1p_10k","pgm":"dictionary_stand_1p"},"transformation":{"data":"N/A","popup":"N/A","suffix":""}}},"surrogate_mean_nch":{"name":"Predicted fatalities, isolated contribution from neighborhood conflict history","description":"Predicted fatalities per type of violence (tv) and level of analysis, given only the neighborhood conflict history in the country at hand.

The prediction is generated by a surrogate or interpretation model that shows how much of the main VIEWS conflict prediction that can be related to variable `splag_1_decay_ged_{tv}_5`; in practice, a regression model with the main prediction as the dependent variable and a flexible function of `splag_1_decay_ged_{tv}_5` as the independent one.

The surrogate model results should be seen as an indication of the relative importance of specific independent variables (predictors), rather than point predictions of their own (they do not sum up to the point predictions from the main ensemble model).

Learn more about the surrogate models.","unit":{"name":"Fatalities","range":"0 - unlimited"},"original":"sc_cm_{tv}_surrogate_nch","type":"Surrogate model","visualization":{"legend":"Predicted fatalities","colormap":"rainbow","dictionary":{"cm":"dictionary_stand_1p_10k","pgm":"dictionary_stand_1p"},"transformation":{"data":"ln","popup":"N/A","suffix":""}}}}