Creating predictions from a model-CodePudding

I am trying to create predictions from my model, but i wish to keep the Country_Name in the predictions if possible. Is there any way this can be done as i'm having no lucking using the standard predict() function

My model is;

mod = gam(gdp_per_capita ~ s(fisheries_production_pc, k = 10, bs = 'cs')   s(food_yield_pc, k = 10, bs = 'cs')  
            s(freshwaster_production_pc, k = 5, bs = 'cs')   s(co2, k = 5, bs = 'cs')   Country_Name, 
            data = economy_df, 
            family = gaussian(link = "log"))

data snipet;

economy_df 
Country_Name year gdp_per_capita Agriculture_GDP_per fisheries_production_pc food_yield_pc freshwaster_production_pc   co2                   
Albania 2018      5287.6637          18.4294792            0.0052701739  1.688718e-03              3.342199e-07  1.782739
Albania 2019      5396.2159          18.3893474            0.0053295312  1.765194e-03              3.342199e-07  1.692248
Albania 2020      5332.1605          19.2644408            0.0059591472  1.835616e-03              3.342199e-07  3.926145
Algeria 2018      4142.0186          11.8742008            0.0028456292  4.622480e-05              2.321186e-07  3.920109
Algeria 2019      3989.6683          12.3362121            0.0024478768  4.105168e-05              2.321186e-07  3.977650
Algeria 2020      3306.8582          14.1347926            0.0019817330  3.467192e-05              2.321186e-07  2.448906
Bosnia 2018      6070.3530           5.8854355            0.0011864874  1.651028e-03              1.206103e-07  6.799183
Bosnia 2019      6119.7624           5.6030922            0.0012912459  1.622146e-03              1.206103e-07  6.382918
Bosnia 2020      6082.3667           6.0844855            0.0012438373  1.844267e-03              1.206103e-07  4.962175
Croatia 2018     15227.5601           2.9570919            0.0220747984  1.725996e-03              1.646345e-07  4.019235
Croatia 2019     15311.7669           2.8687641            0.0209151509  1.760604e-03              1.646345e-07  4.063708
Croatia 2020     14132.4866           3.2165075            0.0230609534  1.727508e-03              1.646345e-07  8.057848
Cyprus 2018     29334.1113           1.7335399            0.0074306923  8.853390e-04              1.740575e-07  6.054175
Cyprus 2019     29206.0762           1.8086052            0.0079922641  2.216217e-03              1.740575e-07  5.998795
Cyprus 2020     27681.5664           1.9308417            0.0071299388  1.961717e-03              1.740575e-07  5.614297
Egypt 2018      2537.1252          11.2250002            0.0199902966  6.887169e-05              7.874128e-07  2.518806
Egypt 2019      3019.0923          11.0489759            0.0203110909  6.022130e-05              7.874128e-07  2.484060
Egypt 2020      3569.2068          11.5676091            0.0196471464  6.046745e-05              7.874128e-07  5.295201

What I'm looking for would look something like this i imagine:

Country_Name      prediction 
Albania            <value> 
Albania            <value>  
Albania            <value>

CodePudding user response：

To check the order of values you can perform a correlation between observed versus predicted values:

cor(economy_df$gdp_per_capita,preds[["fit"]])

Then make new df with your desired columns:

    mypred<-data.frame(Country_Name = 
            economy_df$Country_Name
            ,prediction=preds[["fit"]])
head(mypred)

      Country_Name prediction
1      Albania   5259.356
2      Albania   5382.758
3      Albania   5373.350
4      Algeria   4099.978
5      Algeria   3951.231
6      Algeria   3402.162