{"cells":[{"metadata":{"_uuid":"051d70d956493feee0c6d64651c6a088724dca2a","_execution_state":"idle","trusted":true},"cell_type":"code","source":"library(e1071)\nlibrary(h2o)","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"start_time <- Sys.time()\n\ntrain <- read.csv('/kaggle/input/predict-volcanic-eruptions-ingv-oe/train.csv')\ntrain_set <-data.frame()\n\nfor (i in 1:nrow(train)) {\n  segment_id <- as.character(train[i,1])\n  signals <- read.csv(paste(\"/kaggle/input/predict-volcanic-eruptions-ingv-oe/train/\",segment_id,\".csv\", sep = \"\"))\n  df <- data.frame(segment_id)\n  for (j in 1:10) {\n      col <-signals[,j]\n      df <- cbind(df,a=mean(col) )\n      df <- cbind(df,a=sd(col) )\n      df <- cbind(df,a=var(col) )\n      df <- cbind(df,a=min(col) )\n      df <- cbind(df,a=max(col) )\n      df <- cbind(df,a=median(col) )\n      df <- cbind(df,a=kurtosis (col) )\n      df <- cbind(df,a=skewness(col) )\n      df <- cbind(df,a=mad(col) )\n      \n  }\n  \n  train_set <- rbind(train_set,df)\n}\n\ntrain <- merge(train,train_set, by.x = \"segment_id\")\n\nSys.time() - start_time","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"h2o.init()","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"\ndf_hf <- as.h2o(train)\ny <- \"time_to_eruption\"\nx <- names(train)[3:92]\n\naml <- h2o.automl(x = x, y = y,\n                  training_frame = df_hf,\n                  max_runtime_secs = 600)","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"lb <- aml@leaderboard\nlb","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"start_time <- Sys.time()\nsample_submission <- read.csv('/kaggle/input/predict-volcanic-eruptions-ingv-oe/sample_submission.csv')\ntest_set <-data.frame()\n\nfor (i in 1:nrow(sample_submission)) {\n  segment_id <- as.character(sample_submission[i,1])\n  signals <- read.csv(paste(\"/kaggle/input/predict-volcanic-eruptions-ingv-oe/test/\",segment_id,\".csv\", sep = \"\"))\n  df <- data.frame(segment_id)\n  for (j in 1:10) {\n      col <-signals[,j]\n      df <- cbind(df,a=mean(col) )\n      df <- cbind(df,a=sd(col) )\n      df <- cbind(df,a=var(col) )\n      df <- cbind(df,a=min(col) )\n      df <- cbind(df,a=max(col) )\n      df <- cbind(df,a=median(col) )\n      df <- cbind(df,a=kurtosis (col) )\n      df <- cbind(df,a=skewness(col) )\n      df <- cbind(df,a=mad(col) )\n      \n  }\n  \n  test_set <- rbind(test_set,df)\n}\nSys.time() - start_time\n\n","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"sample_submission <- read.csv('/kaggle/input/predict-volcanic-eruptions-ingv-oe/sample_submission.csv')\ntest <- as.h2o(test_set)\nmodel <- aml@leader\np1 = h2o.predict(model, newdata=test)\np2 <- as.data.frame(p1$predict)\nsample_submission$time_to_eruption=p2$predict\n\nwrite.csv(sample_submission,\"sample_submission.csv\", row.names = FALSE)","execution_count":null,"outputs":[]}],"metadata":{"kernelspec":{"name":"ir","display_name":"R","language":"R"},"language_info":{"name":"R","codemirror_mode":"r","pygments_lexer":"r","mimetype":"text/x-r-source","file_extension":".r","version":"3.6.3"}},"nbformat":4,"nbformat_minor":4}