From 2bf7d63dfcec0fca2f98cbd553cdbf8c16ef3782 Mon Sep 17 00:00:00 2001 From: Loic Guegan Date: Fri, 25 Nov 2022 14:20:55 +0100 Subject: Minor changes --- analysis/days.R | 5 + analysis/figures/combined.pdf | Bin 820342 -> 1235707 bytes analysis/figures/dimension_coverage.pdf | Bin 18266 -> 18248 bytes analysis/figures/dimension_efficiency.pdf | Bin 78811 -> 95369 bytes .../figures/dimension_energy-coverage-policy.pdf | Bin 139129 -> 147139 bytes .../dimension_energy-coverage-wakeupfor.pdf | Bin 139434 -> 147425 bytes analysis/figures/dimension_energy-coverage.pdf | Bin 150220 -> 156877 bytes analysis/figures/dimension_energy.pdf | Bin 111805 -> 127845 bytes analysis/figures/f1_scores_offline.csv | 3 + analysis/figures/knn.csv | 2 + analysis/figures/knn_elbow_NoHintIsFALSE.pdf | Bin 0 -> 5324 bytes analysis/figures/knn_elbow_NoHintIsTRUE.pdf | Bin 0 -> 5290 bytes analysis/figures/random_inputs_NoHintIsFALSE.pdf | Bin 0 -> 159779 bytes analysis/figures/random_inputs_NoHintIsTRUE.pdf | Bin 0 -> 159245 bytes .../figures/sim_dimension_coverage_NO_HINT.pdf | Bin 32846 -> 32982 bytes .../figures/sim_dimension_coverage_WITH_HINT.pdf | Bin 33151 -> 33358 bytes analysis/figures/sim_dimension_energy_NO_HINT.pdf | Bin 37471 -> 38035 bytes .../figures/sim_dimension_energy_WITH_HINT.pdf | Bin 37650 -> 38170 bytes analysis/figures/tree_FALSE.pdf | Bin 0 -> 9593 bytes analysis/figures/tree_TRUE.pdf | Bin 0 -> 7964 bytes analysis/learning.R | 103 ++++++++++++++++++--- 21 files changed, 102 insertions(+), 11 deletions(-) create mode 100644 analysis/figures/f1_scores_offline.csv create mode 100644 analysis/figures/knn.csv create mode 100644 analysis/figures/knn_elbow_NoHintIsFALSE.pdf create mode 100644 analysis/figures/knn_elbow_NoHintIsTRUE.pdf create mode 100644 analysis/figures/random_inputs_NoHintIsFALSE.pdf create mode 100644 analysis/figures/random_inputs_NoHintIsTRUE.pdf create mode 100644 analysis/figures/tree_FALSE.pdf create mode 100644 analysis/figures/tree_TRUE.pdf diff --git a/analysis/days.R b/analysis/days.R index 6ad721e..601f210 100644 --- a/analysis/days.R +++ b/analysis/days.R @@ -1,3 +1,8 @@ +########## INFORMATIONS ########## +# This file is made to study online classification +# So, each pair (wireless,wakeupfor) has its classification models (knn and decision tree) +################################## + library("tidyverse") options(dplyr.summarise.inform = FALSE) library("class") diff --git a/analysis/figures/combined.pdf b/analysis/figures/combined.pdf index d9cbbb8..80f32bd 100644 Binary files a/analysis/figures/combined.pdf and b/analysis/figures/combined.pdf differ diff --git a/analysis/figures/dimension_coverage.pdf b/analysis/figures/dimension_coverage.pdf index a108e34..6761219 100644 Binary files a/analysis/figures/dimension_coverage.pdf and b/analysis/figures/dimension_coverage.pdf differ diff --git a/analysis/figures/dimension_efficiency.pdf b/analysis/figures/dimension_efficiency.pdf index df0eae8..cc4ac9b 100644 Binary files a/analysis/figures/dimension_efficiency.pdf and b/analysis/figures/dimension_efficiency.pdf differ diff --git a/analysis/figures/dimension_energy-coverage-policy.pdf b/analysis/figures/dimension_energy-coverage-policy.pdf index d0c73e8..1bcb5cc 100644 Binary files a/analysis/figures/dimension_energy-coverage-policy.pdf and b/analysis/figures/dimension_energy-coverage-policy.pdf differ diff --git a/analysis/figures/dimension_energy-coverage-wakeupfor.pdf b/analysis/figures/dimension_energy-coverage-wakeupfor.pdf index beacaa9..22d7539 100644 Binary files a/analysis/figures/dimension_energy-coverage-wakeupfor.pdf and b/analysis/figures/dimension_energy-coverage-wakeupfor.pdf differ diff --git a/analysis/figures/dimension_energy-coverage.pdf b/analysis/figures/dimension_energy-coverage.pdf index 9bd624a..02d04ec 100644 Binary files a/analysis/figures/dimension_energy-coverage.pdf and b/analysis/figures/dimension_energy-coverage.pdf differ diff --git a/analysis/figures/dimension_energy.pdf b/analysis/figures/dimension_energy.pdf index 379af96..a2f0c51 100644 Binary files a/analysis/figures/dimension_energy.pdf and b/analysis/figures/dimension_energy.pdf differ diff --git a/analysis/figures/f1_scores_offline.csv b/analysis/figures/f1_scores_offline.csv new file mode 100644 index 0000000..108b671 --- /dev/null +++ b/analysis/figures/f1_scores_offline.csv @@ -0,0 +1,3 @@ +model,f1_baseline,f1_hint,f1_extended,f1_hintandextended +knn,0.88,NA,0.89,0.91 +tree,0.93,NA,0.86,0.92 diff --git a/analysis/figures/knn.csv b/analysis/figures/knn.csv new file mode 100644 index 0000000..3b5cc10 --- /dev/null +++ b/analysis/figures/knn.csv @@ -0,0 +1,2 @@ +f1_baseline +0.905147752632288 diff --git a/analysis/figures/knn_elbow_NoHintIsFALSE.pdf b/analysis/figures/knn_elbow_NoHintIsFALSE.pdf new file mode 100644 index 0000000..31eb996 Binary files /dev/null and b/analysis/figures/knn_elbow_NoHintIsFALSE.pdf differ diff --git a/analysis/figures/knn_elbow_NoHintIsTRUE.pdf b/analysis/figures/knn_elbow_NoHintIsTRUE.pdf new file mode 100644 index 0000000..96072cc Binary files /dev/null and b/analysis/figures/knn_elbow_NoHintIsTRUE.pdf differ diff --git a/analysis/figures/random_inputs_NoHintIsFALSE.pdf b/analysis/figures/random_inputs_NoHintIsFALSE.pdf new file mode 100644 index 0000000..3795adc Binary files /dev/null and b/analysis/figures/random_inputs_NoHintIsFALSE.pdf differ diff --git a/analysis/figures/random_inputs_NoHintIsTRUE.pdf b/analysis/figures/random_inputs_NoHintIsTRUE.pdf new file mode 100644 index 0000000..c5a7506 Binary files /dev/null and b/analysis/figures/random_inputs_NoHintIsTRUE.pdf differ diff --git a/analysis/figures/sim_dimension_coverage_NO_HINT.pdf b/analysis/figures/sim_dimension_coverage_NO_HINT.pdf index a946b86..b9b9519 100644 Binary files a/analysis/figures/sim_dimension_coverage_NO_HINT.pdf and b/analysis/figures/sim_dimension_coverage_NO_HINT.pdf differ diff --git a/analysis/figures/sim_dimension_coverage_WITH_HINT.pdf b/analysis/figures/sim_dimension_coverage_WITH_HINT.pdf index 0f42067..a226cf0 100644 Binary files a/analysis/figures/sim_dimension_coverage_WITH_HINT.pdf and b/analysis/figures/sim_dimension_coverage_WITH_HINT.pdf differ diff --git a/analysis/figures/sim_dimension_energy_NO_HINT.pdf b/analysis/figures/sim_dimension_energy_NO_HINT.pdf index 0f32541..d5ab543 100644 Binary files a/analysis/figures/sim_dimension_energy_NO_HINT.pdf and b/analysis/figures/sim_dimension_energy_NO_HINT.pdf differ diff --git a/analysis/figures/sim_dimension_energy_WITH_HINT.pdf b/analysis/figures/sim_dimension_energy_WITH_HINT.pdf index 7e411e5..590bc80 100644 Binary files a/analysis/figures/sim_dimension_energy_WITH_HINT.pdf and b/analysis/figures/sim_dimension_energy_WITH_HINT.pdf differ diff --git a/analysis/figures/tree_FALSE.pdf b/analysis/figures/tree_FALSE.pdf new file mode 100644 index 0000000..fd9f08d Binary files /dev/null and b/analysis/figures/tree_FALSE.pdf differ diff --git a/analysis/figures/tree_TRUE.pdf b/analysis/figures/tree_TRUE.pdf new file mode 100644 index 0000000..56466fc Binary files /dev/null and b/analysis/figures/tree_TRUE.pdf differ diff --git a/analysis/learning.R b/analysis/learning.R index d21b8cd..ff31d33 100644 --- a/analysis/learning.R +++ b/analysis/learning.R @@ -1,8 +1,10 @@ library("tidyverse") +options(dplyr.summarise.inform = FALSE) library("class") library("rpart") library("rpart.plot") library("viridis") +library("MLmetrics") ## Simulation Parameters: ## simkey {baseline,extended,hint,hintandextended} @@ -24,16 +26,95 @@ nsimkey=4 nsimulations=nseed*nwakeupfor*nwireless*nsimkey # Must be 3200 ## Load data -data=read_csv("../CCGRID2022.csv")%>%distinct() # Note that in the data experiment wireless=="lora",seed==1,wakeupfor==60,simkey=="baseline" is present 2 times in the CSV file -tmp_data_coverage=data%>%group_by(simkey,wireless,wakeupfor,seed)%>%mutate(coverage=sum(nDataRcv))%>%ungroup()%>%filter(isSender==1)%>%select(simkey,wireless,wakeupfor,seed,coverage) -data_seed_isSender=data%>%group_by(simkey,wireless,wakeupfor,seed,isSender)%>%summarize(energy_mean=mean(energy))%>% - left_join(tmp_data_coverage,by=c("simkey","wireless","wakeupfor","seed"))%>% - mutate(efficiency=energy_mean/coverage)%>% - ungroup() +data=suppressMessages(read_csv("../CCGRID2022.csv"))%>%distinct() # Note that in the data experiment wireless=="lora",seed==1,wakeupfor==60,simkey=="baseline" is present 2 times in the CSV file data_seed=data%>%group_by(simkey,wireless,wakeupfor,seed)%>%summarize(energy=sum(energy),coverage=sum(nDataRcv))%>% mutate(efficiency=energy/coverage)%>% ungroup() + +F1_Score2=function(truth, pred){ + result=sapply(c("baseline","extended","hint","hintandextended"),function(c){ + cur_truth=truth[truth==c] + cur_pred=pred[truth==c] + col=paste0("f1_",c) + score=F1_Score(cur_truth,cur_pred) + if(is.nan(score)){score=0} + list(tibble(!!col:=score)) + }) + do.call("cbind",result) +} + +build_models=function(ignore_hint=TRUE){ + ## Prepare data for traning + set.seed(1) # Reproducibility + wireless_map=c("lora"=1,"nbiot"=2) + data_ml=data_seed%>%select(-efficiency,-seed)%>%mutate(wireless=wireless_map[data_seed$wireless]) + if(ignore_hint){ + data_ml=data_ml%>%filter(simkey!="hint") + } + train_set=data_ml%>%sample_frac(0.8) # 80% of the data + test_set=data_ml%>%suppressMessages(anti_join(train_set)) # 20% of the data + + ## KNN predict function + knn_fn=function(inputs){ + as.vector(knn(train=train_set%>%select(-simkey),test=inputs%>%select(-simkey),cl=train_set$simkey,k=10)) + } + + ## Decision tree + tree=rpart( + simkey ~ wireless + wakeupfor + energy + coverage, + data=train_set, + method="class", + minsplit=60, + minbucket=1) + ## Tree predict function + tree_fn=function(inputs){ + as.vector(predict(tree,newdata=inputs%>%select(-simkey),type="class")) + } + + ## Build models + models=list(predict_knn=knn_fn,predict_tree=tree_fn) + + ## Computer performances + perfs=sapply(seq(1,20),function(i){ + ## Prepare data for traning + set.seed(1) # Reproducibility + wireless_map=c("lora"=1,"nbiot"=2) + data_ml=data_seed%>%select(-efficiency,-seed)%>%mutate(wireless=wireless_map[data_seed$wireless]) + if(ignore_hint){ + data_ml=data_ml%>%filter(simkey!="hint") + } + train_set=data_ml%>%sample_frac(0.8) # 80% of the data + test_set=data_ml%>%suppressMessages(anti_join(train_set)) # 20% of the data + + ## KNN + knn_predictions=as.vector(knn(train=train_set%>%select(-simkey),test=test_set%>%select(-simkey),cl=train_set$simkey,k=10)) + ## Decision tree + tree=rpart( + simkey ~ wireless + wakeupfor + energy + coverage, + data=train_set, + method="class", + minsplit=60, + minbucket=1) + tree_predictions=as.vector(predict(tree,newdata=test_set%>%select(-simkey),type="class")) + + ## Prefs + f1_knn=F1_Score2(test_set$simkey,knn_predictions) + f1_tree=F1_Score2(test_set$simkey,tree_predictions) + list(cbind(tibble(model=c("knn","tree")),rbind(f1_knn,f1_tree))) + }) + perfs=do.call("rbind",perfs)%>%mutate_if(is.numeric, ~round(.,digits=2)) + perfs=perfs%>%group_by(model)%>%summarize( + f1_baseline=mean(f1_baseline), + f1_hint=mean(f1_hint), + f1_extended=mean(f1_extended), + f1_hintandextended=mean(f1_hintandextended)) + write.csv(perfs,"figures/f1_scores_offline.csv",quote=FALSE,row.names=FALSE) + + ## Return models + models +} + generate_inputs=function(ignore_hint=FALSE) { ## Prepare data for traning set.seed(1) # Reproducibility @@ -61,7 +142,7 @@ generate_inputs=function(ignore_hint=FALSE) { minbucket=1) tree_predictions=predict(tree,newdata=test_set%>%select(-simkey),type="class") tree_cont_table=table(tree_predictions,test_set$simkey) - tree_accuracy=round((sum(diag(tree_cont_table)/sum(rowSums(tree_cont_table))))*100) + tree_accuracy=(sum(diag(tree_cont_table)/sum(rowSums(tree_cont_table)))) tree_prop_table=round(prop.table(tree_cont_table),digits=2) ## Elbow plot @@ -69,12 +150,12 @@ generate_inputs=function(ignore_hint=FALSE) { knn_predictions=knn(train=train_set%>%select(-simkey),test=test_set%>%select(-simkey),cl=train_set$simkey,k=kvalue) ## KNN analysis knn_cont_table=table(knn_predictions,test_set$simkey) - knn_accuracy=round((sum(diag(knn_cont_table)/sum(rowSums(knn_cont_table))))*100) + knn_accuracy=(sum(diag(knn_cont_table)/sum(rowSums(knn_cont_table)))) knn_prop_table=round(prop.table(knn_cont_table),digits=2) tibble(k=kvalue,accuracy=knn_accuracy) }) elbow_data=do.call("rbind",elbow_data) - ggplot(data=elbow_data,aes(k,accuracy))+geom_point()+geom_line()+ggtitle(paste("K-elbow for with NoHint to",as.character(ignore_hint))) + ggplot(data=elbow_data,aes(k,accuracy))+geom_point()+geom_line()+ggtitle(paste("K-elbow for with NoHint to",as.character(ignore_hint)))+ylim(c(0,1)) ggsave(paste0("figures/knn_elbow_NoHintIs",as.character(ignore_hint),".pdf")) ## Prints @@ -126,5 +207,5 @@ generate_inputs=function(ignore_hint=FALSE) { } ## Generate inputs -generate_inputs(FALSE) -generate_inputs(TRUE) +#generate_inputs(FALSE) +#generate_inputs(TRUE) -- cgit v1.2.3