**：
package mllibimport org.apache.log4j.
import org.apache.spark.
import scala.collection.mutable.hashmap
/***
隨機梯度下降演算法
* created by
汪本成
on 2016/8/5.
*/object sgd
data
} //
假設a=0
var
a: double = 0
//設定步進係數
var
b: double = 0.1
//設定迭代公式
def
sgd(x: double, y: double) =
def
main(args: array[string]) )
//顯示結果
println("
最終結果a為
" + a)
}}
執行結果：
sgdusing spark's default log4j profile: org/apache/spark/log4j-defaults.properties
slf4j: class path contains multiple slf4j bindings.
slf4j: found binding in [jar:file:/g:/home/download/spark-1.6.1-bin-hadoop2.6/lib/spark-assembly-1.6.1-hadoop2.6.0.jar!/org/slf4j/impl/staticloggerbinder.class]
slf4j: found binding in [jar:file:/g:/home/download/spark-1.6.1-bin-hadoop2.6/lib/spark-examples-1.6.1-hadoop2.6.0.jar!/org/slf4j/impl/staticloggerbinder.class]
slf4j: see for an explanation.
slf4j: actual binding is of type [org.slf4j.impl.log4jlogge***ctory]
16/08/05 00:48:28 info slf4jlogger: slf4jlogger started
16/08/05 00:48:28 info remoting: starting remoting
16/08/05 00:48:28 info remoting: remoting started; listening on addresses :[akka.tcp:
data:
(23,46)
(50,100)
(32,64)
(41,82)
(17,34)
(8,16)
(35,70)
(44,88)
(26,52)
(11,22)
(29,58)
(38,76)
(47,94)
(20,40)
(2,4)
(5,10)
(14,28)
(46,92)
(40,80)
(49,98)
(4,8)
(13,26)
(22,44)
(31,62)
(16,32)
(7,14)
(43,86)
(25,50)
(34,68)
(10,20)
(37,74)
(1,2)
(19,38)
(28,56)
(45,90)
(27,54)
(36,72)
(18,36)
(9,18)
(21,42)
(48,96)
(3,6)
(12,24)
(30,60)
(39,78)
(15,30)
(42,84)
(24,48)
(6,12)
(33,66)
result:
1:0.0(23,46)
2:4.6000000000000005(50,100)
3:-8.400000000000002(32,64)
4:24.880000000000006(41,82)
5:-68.92800000000003(17,34)
6:51.649600000000035(8,16)
7:11.929920000000003(35,70)
8:-22.82480000000001(44,88)
9:86.40432000000006(26,52)
10:-133.04691200000013(11,22)
11:15.504691199999996(29,58)
12:-23.65891328(38,76)
13:73.84495718400001(47,94)
14:-263.82634158080003(20,40)
15:267.82634158080003(2,4)
16:214.66107326464004(5,10)
17:108.33053663232002(14,28)
18:-40.53221465292802(46,92)
19:155.1159727505409(40,80)
20:-457.3479182516227(49,98)
21:1793.4568811813288(4,8)
22:1076.8741287087973(13,26)
23:-320.46223861263934(22,44)
24:388.95468633516725(31,62)
25:-810.6048413038511(16,32)
26:489.56290478231085(7,14)
27:148.2688714346932(43,86)
28:-480.6872757344877(25,50)
29:726.0309136017315(34,68)
30:-1735.6741926441557(10,20)
31:2.0000000000002274(37,74)
32:1.999999999999386(1,2)
33:1.9999999999994476(19,38)
34:2.000000000000497(28,56)
35:1.9999999999991056(45,90)
36:2.00000000000313(27,54)
37:1.9999999999946787(36,72)
38:2.000000000013835(18,36)
39:1.999999999988932(9,18)
40:1.999999999998893(21,42)
41:2.0000000000012172(48,96)
42:1.9999999999953737(3,6)
43:1.9999999999967615(12,24)
44:2.000000000000648(30,60)
45:1.999999999998704(39,78)
46:2.0000000000037588(15,30)
47:1.9999999999981206(42,84)
48:2.0000000000060134(24,48)
49:1.999999999991581(6,12)
50:1.9999999999966325(33,66)
最終結果a為 2.0000000000077454
16/08/05 00:48:28 info remoteactorrefprovider$remotingterminator: shutting down remote daemon.
process finished with exit code 0
分析：
當α為0.1的時候，一般30次計算就計算出來了；如果是0.5，一般15次計算就有正確結果 。如果是1，則50次都沒有結果
隨機打亂 隨機發牌
一.產生乙個隨機數 1.new random nextint 10 0,10 的隨機數 2.二.如何打亂乙個陣列或隨機排序，除了 random之外collections.shuffle非常方便，且效率高。可以實現隨機打亂列表的功能，實現把員工 順序 打亂等，安全傳輸方面，比如傳送加密前，把資料隨機打...
隨機森林隨機 三
2018年7月22日 陣雨 感冒發燒喉嚨疼，只因為一杯正常冰的奶蓋！以後得少喝加冰的東西了.前面說的是整合學習的兩種形式，這跟隨機森林有啥關係呢？隨機森林是bagging的乙個擴充套件變體.簡單的來說，當bagging演算法中的base learner為決策樹並在訓練決策樹的過程中加入隨機屬性選擇，...
隨機檔案與隨機函式
1.dev random 與 dev urandom 特殊檔案 有些系統會提供兩種隨機偽裝置 dev random 與 dev urandom。這兩個裝置的差別，在 dev random 會一直封鎖，直到系統所產生的隨機數已充分夠用，所以它可以確保高品質的隨機數。相對地，dev urandom 不會...