Group_randomfrost_lin_h2o.R
daitu
Sat Jun 25 15:58:08 2016
## 2016年暑期课程设计####
## 问题：Grupo Bimbo Inventory Demand
## 宾堡集团的库存需求
## 最大限度地提高销售和最大限度地减少烘焙食品的退回
## Daitu
## start:2016.06.22
## 参考借鉴kaggle上的公开程序
## 使用使用随机森林模型进行预测
##设置工作文件夹
setwd("/Users/Daitu/数据分析/kaggle/Grupo Bimbo")
getwd()## [1] "/Users/daitu/数据分析/kaggle/Grupo Bimbo"## 设置集群 ####
print(paste("Set up Cluster",Sys.time()))## [1] "Set up Cluster 2016-06-25 15:58:08"library(h2o) # R API is just a library## Warning: package 'h2o' was built under R version 3.2.5## Loading required package: statmod##
## ----------------------------------------------------------------------
##
## Your next step is to start H2O:
## > h2o.init()
##
## For H2O package documentation, ask for help:
## > ??h2o
##
## After starting H2O, you can use the Web UI at http://localhost:54321
## For more information visit http://docs.h2o.ai
##
## ----------------------------------------------------------------------##
## Attaching package: 'h2o'## The following objects are masked from 'package:stats':
##
## sd, var## The following objects are masked from 'package:base':
##
## &&, %*%, %in%, ||, apply, as.factor, as.numeric, colnames,
## colnames<-, ifelse, is.character, is.factor, is.numeric, log,
## log10, log1p, log2, round, signif, trunc## 启动一个集群; 定义为4核同时计算；
h2o.init(nthreads=6,max_mem_size='12G') ##
## H2O is not running yet, starting it now...
##
## Note: In case of errors look at the following log files:
## /var/folders/bh/xgh997m97vl8yvm3sxvwfr5r0000gn/T//RtmpC56yU8/h2o_daitu_started_from_r.out
## /var/folders/bh/xgh997m97vl8yvm3sxvwfr5r0000gn/T//RtmpC56yU8/h2o_daitu_started_from_r.err
##
##
## Starting H2O JVM and connecting: .. Connection successful!
##
## R is connected to the H2O cluster:
## H2O cluster uptime: 1 seconds 606 milliseconds
## H2O cluster version: 3.8.2.6
## H2O cluster name: H2O_started_from_R_daitu_kqf986
## H2O cluster total nodes: 1
## H2O cluster total memory: 10.67 GB
## H2O cluster total cores: 8
## H2O cluster allowed cores: 6
## H2O cluster healthy: TRUE
## H2O Connection ip: localhost
## H2O Connection port: 54321
## H2O Connection proxy: NA
## R Version: R version 3.2.3 (2015-12-10)## 加载数据####
print(paste("加载数据",Sys.time()))## [1] "加载数据 2016-06-25 15:58:11"## 读取整个训练数据，使用所有的核
system.time({
train<-h2o.uploadFile("train.csv",destination_frame = "train.hex")
})##
|
| | 0%
|
|=================================================================| 100%## user system elapsed
## 0.386 2.907 19.656train[1:5,] ## 查看训练集的前几行## Semana Agencia_ID Canal_ID Ruta_SAK Cliente_ID Producto_ID Venta_uni_hoy
## 1 3 1110 7 3301 15766 1212 3
## 2 3 1110 7 3301 15766 1216 4
## 3 3 1110 7 3301 15766 1238 4
## 4 3 1110 7 3301 15766 1240 4
## 5 3 1110 7 3301 15766 1242 3
## Venta_hoy Dev_uni_proxima Dev_proxima Demanda_uni_equil
## 1 25.14 0 0 3
## 2 33.52 0 0 4
## 3 39.32 0 0 4
## 4 33.52 0 0 4
## 5 22.92 0 0 3
##
## [5 rows x 11 columns]## 将训练集
train$target<- log(train$Demanda_uni_equil+1)
train[1:5,]## Semana Agencia_ID Canal_ID Ruta_SAK Cliente_ID Producto_ID Venta_uni_hoy
## 1 3 1110 7 3301 15766 1212 3
## 2 3 1110 7 3301 15766 1216 4
## 3 3 1110 7 3301 15766 1238 4
## 4 3 1110 7 3301 15766 1240 4
## 5 3 1110 7 3301 15766 1242 3
## Venta_hoy Dev_uni_proxima Dev_proxima Demanda_uni_equil target
## 1 25.14 0 0 3 1.386294
## 2 33.52 0 0 4 1.609438
## 3 39.32 0 0 4 1.609438
## 4 33.52 0 0 4 1.609438
## 5 22.92 0 0 3 1.386294
##
## [5 rows x 12 columns]dim(train)## [1] 74180464 12h2o.median(train$target)## [1] 1.386294summary(train$target)## Warning in summary.H2OFrame(train$target): Approximated quantiles
## computed! If you are interested in exact quantiles, please pass the
## `exact_quantiles=TRUE` parameter.## target
## Min. :0.000
## 1st Qu.:1.090
## Median :1.380
## Mean :1.603
## 3rd Qu.:1.942
## Max. :8.517## 数据分区
print(paste("数据分区",Sys.time()))## [1] "数据分区 2016-06-25 15:58:36"## 这个模型将会把数据分为3个部分，根据星期数据进行分区：
## one to generate product averages, a second to fit a model, and a third to evaluate the model
## 第一个数据用来生成产品均值，第二部分数据用来拟合一个模型，第三部分数据用来计算模型
dev<-train[train$Semana <= 5,] ## gets Semana 3,4,5
dim(dev)## [1] 32790197 12val<-train[train$Semana > 4 & train$Semana <= 8,] ## gets Semana 5,6, 7,8
val[1:5,]## Semana Agencia_ID Canal_ID Ruta_SAK Cliente_ID Producto_ID Venta_uni_hoy
## 1 5 1110 7 3301 15766 1212 5
## 2 5 1110 7 3301 15766 1216 3
## 3 5 1110 7 3301 15766 1220 3
## 4 5 1110 7 3301 15766 1238 1
## 5 5 1110 7 3301 15766 1242 2
## Venta_hoy Dev_uni_proxima Dev_proxima Demanda_uni_equil target
## 1 41.90 0 0 5 1.7917595
## 2 25.14 0 0 3 1.3862944
## 3 22.92 0 0 3 1.3862944
## 4 9.83 0 0 1 0.6931472
## 5 15.28 0 0 2 1.0986123
##
## [5 rows x 12 columns]dim(val)## [1] 41596951 12final<-train[train$Semana >=8,] ## gets Semana 8,9
final[1:5,]## Semana Agencia_ID Canal_ID Ruta_SAK Cliente_ID Producto_ID Venta_uni_hoy
## 1 8 1110 7 3301 15766 1212 4
## 2 8 1110 7 3301 15766 1216 5
## 3 8 1110 7 3301 15766 1220 1
## 4 8 1110 7 3301 15766 1238 3
## 5 8 1110 7 3301 15766 1240 2
## Venta_hoy Dev_uni_proxima Dev_proxima Demanda_uni_equil target
## 1 33.52 0 0 4 1.6094379
## 2 41.90 0 0 5 1.7917595
## 3 7.64 0 0 1 0.6931472
## 4 29.49 0 0 3 1.3862944
## 5 16.76 0 0 2 1.0986123
##
## [5 rows x 12 columns]dim(final)## [1] 20815581 12## 模型：RandomForest####
print(paste("Model: Product Groups & randomForest",Sys.time()))## [1] "Model: Product Groups & randomForest 2016-06-25 15:59:03"## 使用测试集中用来预测的字段变量进行预测，剔除ID和星期，
predictors<-c("Agencia_ID","Canal_ID","Ruta_SAK","Cliente_ID","Producto_ID")
## first part of model: use product averages, created on the dev set
## this is the only time we will use the dev set
## 模型的第一部分：使用产品的均值，在dev数据集上创建
## 这是dev数据集的唯一的一次使用
groups<-h2o.group_by(data=dev,by=c("Producto_ID","Canal_ID"),mean("target"))
groups[1:5,]## Producto_ID Canal_ID mean_target
## 1 41 7 4.357809
## 2 53 4 5.852552
## 3 72 1 1.644182
## 4 72 6 2.378727
## 5 72 7 2.608634
##
## [5 rows x 3 columns]h2o.median(groups$mean_target)## [1] 2.118703dim(groups)## [1] 4657 3## apply groups back into dev and validation data sets as "mean_target"
## if there are NAs for this (new products), use a constant; used median of entire train target
## 使用分组后的数据集dev，生成新的确认数据（val）
## 如果数据集中有NAS（代表新的产品），使用中位数进行代替。
newVal<-h2o.merge(x=val,y=groups,all.x = T)
newVal[1:5,]## Canal_ID Producto_ID Semana Ruta_SAK Cliente_ID Agencia_ID Venta_uni_hoy
## 1 7 1212 5 3301 15766 1110 5
## 2 7 1216 5 3301 15766 1110 3
## 3 7 1220 5 3301 15766 1110 3
## 4 7 1238 5 3301 15766 1110 1
## 5 7 1242 5 3301 15766 1110 2
## Venta_hoy Dev_uni_proxima Dev_proxima Demanda_uni_equil target
## 1 41.90 0 0 5 1.7917595
## 2 25.14 0 0 3 1.3862944
## 3 22.92 0 0 3 1.3862944
## 4 9.83 0 0 1 0.6931472
## 5 15.28 0 0 2 1.0986123
## mean_target
## 1 1.535514
## 2 1.537746
## 3 1.532841
## 4 1.634883
## 5 1.780982
##
## [5 rows x 13 columns]newVal$mean_target[is.na(newVal$mean_target)]<-h2o.median(groups$mean_target)
newVal[1:10,]## Canal_ID Producto_ID Semana Ruta_SAK Cliente_ID Agencia_ID Venta_uni_hoy
## 1 7 1212 5 3301 15766 1110 5
## 2 7 1216 5 3301 15766 1110 3
## 3 7 1220 5 3301 15766 1110 3
## 4 7 1238 5 3301 15766 1110 1
## 5 7 1242 5 3301 15766 1110 2
## 6 7 1250 5 3301 15766 1110 8
## Venta_hoy Dev_uni_proxima Dev_proxima Demanda_uni_equil target
## 1 41.90 0 0 5 1.7917595
## 2 25.14 0 0 3 1.3862944
## 3 22.92 0 0 3 1.3862944
## 4 9.83 0 0 1 0.6931472
## 5 15.28 0 0 2 1.0986123
## 6 61.12 0 0 8 2.1972246
## mean_target
## 1 1.535514
## 2 1.537746
## 3 1.532841
## 4 1.634883
## 5 1.780982
## 6 1.994878
##
## [10 rows x 13 columns]dim(newVal)## [1] 41596951 13newFinal<-h2o.merge(x=final,y=groups,all.x = T)
newFinal[1:5,]## Canal_ID Producto_ID Semana Ruta_SAK Cliente_ID Agencia_ID Venta_uni_hoy
## 1 7 1212 8 3301 15766 1110 4
## 2 7 1216 8 3301 15766 1110 5
## 3 7 1220 8 3301 15766 1110 1
## 4 7 1238 8 3301 15766 1110 3
## 5 7 1240 8 3301 15766 1110 2
## Venta_hoy Dev_uni_proxima Dev_proxima Demanda_uni_equil target
## 1 33.52 0 0 4 1.6094379
## 2 41.90 0 0 5 1.7917595
## 3 7.64 0 0 1 0.6931472
## 4 29.49 0 0 3 1.3862944
## 5 16.76 0 0 2 1.0986123
## mean_target
## 1 1.535514
## 2 1.537746
## 3 1.532841
## 4 1.634883
## 5 1.836763
##
## [5 rows x 13 columns]newFinal$mean_target[is.na(newFinal$mean_target)]<-h2o.median(groups$mean_target)
newFinal[1:5,]## Canal_ID Producto_ID Semana Ruta_SAK Cliente_ID Agencia_ID Venta_uni_hoy
## 1 7 1212 8 3301 15766 1110 4
## 2 7 1216 8 3301 15766 1110 5
## 3 7 1220 8 3301 15766 1110 1
## 4 7 1238 8 3301 15766 1110 3
## 5 7 1240 8 3301 15766 1110 2
## Venta_hoy Dev_uni_proxima Dev_proxima Demanda_uni_equil target
## 1 33.52 0 0 4 1.6094379
## 2 41.90 0 0 5 1.7917595
## 3 7.64 0 0 1 0.6931472
## 4 29.49 0 0 3 1.3862944
## 5 16.76 0 0 2 1.0986123
## mean_target
## 1 1.535514
## 2 1.537746
## 3 1.532841
## 4 1.634883
## 5 1.836763
##
## [5 rows x 13 columns]## 训练 randomForest; 使用参数以保持整体运行时间在20分钟内
## this model is fit on Semana 6 & 7 & 8, and evaluated on Semana 9.
rf<-h2o.randomForest(
x=predictors, ## 建立模型的预测变量
y="target", ## target: using the logged variable created earlier
training_frame = newVal, ## H2O frame holding the training data
validation_frame = newFinal, ## extra holdout piece for three layer modeling
model_id="randomForest1", ## internal H2O name for model
ntrees = 200, ## 使用100棵树建立模型
sample_rate = 0.8, ## use 80% of the rows each scoring round
weights_column = "mean_target"
)##
|
| | 0%
|
|= | 1%
|
|= | 2%
|
|== | 2%
|
|== | 3%
|
|== | 4%
|
|=== | 4%
|
|=== | 5%
|
|==== | 6%
|
|===== | 7%
|
|===== | 8%
|
|====== | 8%
|
|====== | 9%
|
|====== | 10%
|
|======= | 10%
|
|======= | 11%
|
|======= | 12%
|
|======== | 12%
|
|======== | 13%
|
|========= | 14%
|
|========== | 15%
|
|========== | 16%
|
|=========== | 16%
|
|=========== | 17%
|
|=========== | 18%
|
|============ | 18%
|
|============ | 19%
|
|============= | 20%
|
|============== | 21%
|
|============== | 22%
|
|=============== | 22%
|
|=============== | 23%
|
|=============== | 24%
|
|================ | 24%
|
|================ | 25%
|
|================= | 26%
|
|================== | 27%
|
|================== | 28%
|
|=================== | 28%
|
|=================== | 29%
|
|=================== | 30%
|
|==================== | 30%
|
|==================== | 31%
|
|==================== | 32%
|
|===================== | 32%
|
|===================== | 33%
|
|====================== | 34%
|
|======================= | 35%
|
|======================= | 36%
|
|======================== | 36%
|
|======================== | 37%
|
|======================== | 38%
|
|========================= | 38%
|
|========================= | 39%
|
|========================== | 40%
|
|=========================== | 41%
|
|=========================== | 42%
|
|============================ | 42%
|
|============================ | 43%
|
|============================ | 44%
|
|============================= | 44%
|
|============================= | 45%
|
|============================== | 46%
|
|=============================== | 47%
|
|=============================== | 48%
|
|================================ | 48%
|
|================================ | 49%
|
|================================ | 50%
|
|================================= | 50%
|
|================================= | 51%
|
|================================= | 52%
|
|================================== | 52%
|
|================================== | 53%
|
|=================================== | 54%
|
|=================================== | 55%
|
|==================================== | 55%
|
|==================================== | 56%
|
|===================================== | 56%
|
|===================================== | 57%
|
|====================================== | 58%
|
|====================================== | 59%
|
|======================================= | 60%
|
|======================================== | 61%
|
|======================================== | 62%
|
|========================================= | 62%
|
|========================================= | 63%
|
|========================================= | 64%
|
|========================================== | 64%
|
|========================================== | 65%
|
|=========================================== | 66%
|
|============================================ | 67%
|
|============================================ | 68%
|
|============================================= | 68%
|
|============================================= | 69%
|
|============================================= | 70%
|
|============================================== | 70%
|
|============================================== | 71%
|
|============================================== | 72%
|
|=============================================== | 72%
|
|=============================================== | 73%
|
|================================================ | 74%
|
|================================================= | 75%
|
|================================================= | 76%
|
|================================================== | 76%
|
|================================================== | 77%
|
|================================================== | 78%
|
|=================================================== | 78%
|
|=================================================== | 79%
|
|==================================================== | 80%
|
|===================================================== | 81%
|
|===================================================== | 82%
|
|====================================================== | 82%
|
|====================================================== | 83%
|
|====================================================== | 84%
|
|======================================================= | 84%
|
|======================================================= | 85%
|
|======================================================== | 86%
|
|========================================================= | 87%
|
|========================================================= | 88%
|
|========================================================== | 88%
|
|========================================================== | 89%
|
|========================================================== | 90%
|
|=========================================================== | 90%
|
|=========================================================== | 91%
|
|=========================================================== | 92%
|
|============================================================ | 92%
|
|============================================================ | 93%
|
|============================================================= | 94%
|
|============================================================== | 95%
|
|============================================================== | 96%
|
|=============================================================== | 96%
|
|=============================================================== | 97%
|
|=============================================================== | 98%
|
|================================================================ | 98%
|
|================================================================ | 99%
|
|=================================================================| 100%## 查看模型
summary(rf)## Model Details:
## ==============
##
## H2ORegressionModel: drf
## Model Key: randomForest1
## Model Summary:
## number_of_trees model_size_in_bytes min_depth max_depth mean_depth
## 1 200 60660176 20 20 20.00000
## min_leaves max_leaves mean_leaves
## 1 368 104998 26256.45500
##
## H2ORegressionMetrics: drf
## ** Reported on training data. **
## Description: Metrics reported on Out-Of-Bag training samples
##
## MSE: 0.5565291
## R2 : 0.3818045
## Mean Residual Deviance : 0.5565291
##
##
## H2ORegressionMetrics: drf
## ** Reported on validation data. **
##
## MSE: 0.557135
## R2 : 0.3774114
## Mean Residual Deviance : 0.557135
##
##
##
##
## Scoring History:
## timestamp duration number_of_trees training_MSE
## 1 2016-06-25 15:59:39 0.007 sec 0
## 2 2016-06-25 15:59:59 20.118 sec 1 0.54708
## 3 2016-06-25 16:00:40 1 min 0.640 sec 4 0.58149
## 4 2016-06-25 16:02:17 2 min 38.377 sec 11 0.60043
## 5 2016-06-25 16:06:22 6 min 43.491 sec 26 0.58038
## 6 2016-06-25 16:16:59 17 min 19.760 sec 56 0.55842
## 7 2016-06-25 16:42:33 42 min 54.495 sec 139 0.55277
## 8 2016-06-25 17:04:46 1:05:06.746 200 0.55653
## training_deviance validation_MSE validation_deviance
## 1
## 2 0.54708 0.55148 0.55148
## 3 0.58149 0.53614 0.53614
## 4 0.60043 0.56081 0.56081
## 5 0.58038 0.56545 0.56545
## 6 0.55842 0.55364 0.55364
## 7 0.55277 0.55272 0.55272
## 8 0.55653 0.55713 0.55713
##
## Variable Importances: (Extract with `h2o.varimp`)
## =================================================
##
## Variable Importances:
## variable relative_importance scaled_importance percentage
## 1 Producto_ID 1136510336.000000 1.000000 0.385112
## 2 Canal_ID 985370368.000000 0.867014 0.333897
## 3 Ruta_SAK 581313536.000000 0.511490 0.196981
## 4 Agencia_ID 153383392.000000 0.134960 0.051975
## 5 Cliente_ID 94541512.000000 0.083186 0.032036# 删除不再需要的较大的数据集
h2o.rm(train)
h2o.rm(dev)
h2o.rm(val)
h2o.rm(newVal)
## 进行预测#####
print(paste("Create Predictions",Sys.time()))## [1] "Create Predictions 2016-06-25 17:13:16"## 加载测试集
test<-h2o.uploadFile("test.csv",destination_frame = "test.hex")##
|
| | 0%
|
|=================================================================| 100%test[1:5,] ## 查看测试集的前几行数据## id Semana Agencia_ID Canal_ID Ruta_SAK Cliente_ID Producto_ID
## 1 0 11 4037 1 2209 4639078 35305
## 2 1 11 2237 1 1226 4705135 1238
## 3 2 10 2045 1 2831 4549769 32940
## 4 3 11 1227 1 4448 4717855 43066
## 5 4 11 1219 1 1130 966351 1277
##
## [5 rows x 7 columns]## merge in the offset column, just as with val and final
newTest<-h2o.merge(x=test,y=groups,all.x = T)
newTest[1:5,]## Canal_ID Producto_ID Agencia_ID id Ruta_SAK Cliente_ID Semana
## 1 1 35305 4037 0 2209 4639078 11
## 2 1 1238 2237 1 1226 4705135 11
## 3 1 32940 2045 2 2831 4549769 10
## 4 1 43066 1227 3 4448 4717855 11
## 5 1 1277 1219 4 1130 966351 11
## mean_target
## 1 NaN
## 2 1.216841
## 3 1.433608
## 4 1.043424
## 5 NaN
##
## [5 rows x 8 columns]dim(newTest)## [1] 6999251 8newTest$mean_target[is.na(newTest$mean_target)]<-h2o.median(groups$mean_target)
newTest[1:5,]## Canal_ID Producto_ID Agencia_ID id Ruta_SAK Cliente_ID Semana
## 1 1 35305 4037 0 2209 4639078 11
## 2 1 1238 2237 1 1226 4705135 11
## 3 1 32940 2045 2 2831 4549769 10
## 4 1 43066 1227 3 4448 4717855 11
## 5 1 1277 1219 4 1130 966351 11
## mean_target
## 1 2.118703
## 2 1.216841
## 3 1.433608
## 4 1.043424
## 5 2.118703
##
## [5 rows x 8 columns]p<-h2o.predict(rf,newTest)##
|
| | 0%
|
|= | 2%
|
|=== | 5%
|
|====== | 10%
|
|======== | 12%
|
|============= | 20%
|
|============== | 22%
|
|==================== | 30%
|
|===================== | 32%
|
|======================== | 37%
|
|========================== | 40%
|
|=========================== | 42%
|
|================================ | 50%
|
|================================== | 52%
|
|====================================== | 58%
|
|======================================= | 60%
|
|======================================== | 62%
|
|=========================================== | 67%
|
|============================================ | 68%
|
|============================================== | 70%
|
|=============================================== | 72%
|
|================================================ | 73%
|
|================================================== | 77%
|
|==================================================== | 80%
|
|===================================================== | 82%
|
|======================================================== | 87%
|
|========================================================== | 90%
|
|============================================================ | 92%
|
|============================================================== | 95%
|
|=================================================================| 100%p <- exp(p)-1
dim(p)## [1] 6999251 1summary(p)## Warning in summary.H2OFrame(p): Approximated quantiles computed! If you
## are interested in exact quantiles, please pass the `exact_quantiles=TRUE`
## parameter.## C1
## Min. : 0.000
## 1st Qu.: 3.471
## Median : 4.084
## Mean : 4.967
## 3rd Qu.: 4.900
## Max. :102.092## 对预测值四舍五入
p <- round(p)
p[p<0]<- 0
sum(p == 0)## [1] 120## 预测值的分布
summary(p)## Warning in summary.H2OFrame(p): Approximated quantiles computed! If you
## are interested in exact quantiles, please pass the `exact_quantiles=TRUE`
## parameter.## C1
## Min. : 0.000
## 1st Qu.: 4.000
## Median : 4.000
## Mean : 4.965
## 3rd Qu.: 5.000
## Max. :102.000h2o.hist(p[p<200],breaks = "FD")## Warning in histo$counts/sum(histo$counts) * 1/diff(histo$breaks): 长的对象
## 长度不是短的对象长度的整倍数
h2o.hist(p[p<500],breaks = "FD")## Warning in histo$counts/sum(histo$counts) * 1/diff(histo$breaks): 长的对象
## 长度不是短的对象长度的整倍数
h2o.hist(p,breaks = "FD" )## Warning in histo$counts/sum(histo$counts) * 1/diff(histo$breaks): 长的对象
## 长度不是短的对象长度的整倍数
## 创建提交文件#####
print(paste("Create Submission",Sys.time()))## [1] "Create Submission 2016-06-25 17:16:35"submissionFrame<-h2o.cbind(test$id,p)
dim(submissionFrame)## [1] 6999251 2colnames(submissionFrame)<-c("id","Demanda_uni_equil")
submissionFrame[1:10,]## id Demanda_uni_equil
## 1 0 4
## 2 1 3
## 3 2 4
## 4 3 4
## 5 4 7
## 6 5 5
##
## [10 rows x 2 columns]h2o.exportFile(submissionFrame,path="sh2o_rf.csv") ## 输出文件##
|
| | 0%
|
|=== | 5%
|
|====== | 10%
|
|=========== | 17%
|
|=============== | 23%
|
|==================== | 30%
|
|======================== | 37%
|
|=========================== | 42%
|
|=============================== | 48%
|
|==================================== | 55%
|
|======================================== | 62%
|
|=========================================== | 67%
|
|================================================ | 73%
|
|==================================================== | 80%
|
|======================================================== | 87%
|
|============================================================= | 93%
|
|=================================================================| 100%