Commit e8e6082191aee1985018555da1736a16a53eb223
1 parent
e4f0b08233
Exists in
master
...
Showing 1 changed file with 4 additions and 4 deletions Side-by-side Diff
app/com/piki_ds/ver2ggh/gghScore.scala
View file @
e8e6082
| ... | ... | @@ -58,9 +58,9 @@ |
| 58 | 58 | map(x => (x.getAs[Long]("contents_id"), x.getAs[Long]("cardSize"), x.getAs[Long]("cardSize")/5)).toDF("cid1","cardSize","sizeGroup"). |
| 59 | 59 | selectExpr("*","if(sizeGroup>19, 20, sizeGroup) as cardGroup").drop("sizeGroup") |
| 60 | 60 | /// kpi3 지표로 부터 카드 그룹별 평균 컨텐츠 소비시간 계산 |
| 61 | - val kpi3 = Util.readDashboardTable("kpi3","*","{*}","*", "*").selectExpr("cid","udate","appView","consumeTime","numberOfCard").dropDuplicates(Seq("cid")). | |
| 62 | - join(cardSize, column("cid")===cardSize("cid1")).drop(cardSize("cid1")).where("numberOfCard is not null").cache | |
| 63 | - val cardGroupConsume = kpi3.where("consumeTime > 10.0").groupBy("cardGroup").agg(expr("avg(consumeTime) * 1000 as cardGroupConTime")).selectExpr("cardGroup as cardGroup1", "cardGroupConTime") | |
| 61 | + //val kpi3 = Util.readDashboardTable("kpi3","*","{*}","*", "*").selectExpr("cid","udate","appView","consumeTime","numberOfCard").dropDuplicates(Seq("cid")). | |
| 62 | + // join(cardSize, column("cid")===cardSize("cid1")).drop(cardSize("cid1")).where("numberOfCard is not null").cache | |
| 63 | + //val cardGroupConsume = kpi3.where("consumeTime > 10.0").groupBy("cardGroup").agg(expr("avg(consumeTime) * 1000 as cardGroupConTime")).selectExpr("cardGroup as cardGroup1", "cardGroupConTime") | |
| 64 | 64 | // kpi3.map(x => s"${x(0)}|${x(1)}|${x(2)}|${x(3)}|${x(4)}|${x(5)}|${x(6)}").coalesce(1, shuffle = true).saveAsTextFile(s"hdfs://pikinn/user/evan/Features/kpi3") |
| 65 | 65 | // cardGroupConsume.stat.corr("cardGroup1","cardGroupConTime") cardGroup 과 소비시간평균과의 상관관계 0.89 |
| 66 | 66 | |
| ... | ... | @@ -156,7 +156,7 @@ |
| 156 | 156 | val gghScaled = gghVer3.withColumn("gghScaled", (column("ggh") - gghMean) / gghStd).selectExpr("*", "1000 / (1 + exp(-gghScaled)) as scaledGgh").drop("gghScaled") |
| 157 | 157 | |
| 158 | 158 | gghScaled.map{x => |
| 159 | - s"${x(0)},${x(5)}" | |
| 159 | + s"${x(0)},${x.getAs[Double]("gghScaled").toInt}" | |
| 160 | 160 | }.saveAsTextFile(s"hdfs://pikinn/preprocess/timelineScore/content/ggh/$saveDay") |
| 161 | 161 | |
| 162 | 162 | /*//////// CTR and CTR Time 계산 |