GO语言利用K近邻算法实现小说鉴黄

usuage:
  go run knn.go --file="data.txt"

关键是向量点的选择和阈值的判定
样本数据来自国家新闻出版总署发布通知公布的《40部淫秽色情网络小说名单》

package main
 
import (
  "bufio"
  "flag"
  "fmt"
  "io"
  "log"
  "math"
  "os"
  "path"
  "path/filepath"
)
 
var debug bool = false
var data_dir string = "./moyan"    //文件存放目录
var limen float64 = 0.1159203888322267 //阈值
 
const (
  min_hanzi rune = 0x3400
  max_hanzi rune = 0x9fbb
)
 
var labels []rune = []rune{
  0x817f, 0x80f8, 0x4e73, 0x81c0,
  0x5c41, 0x80a1, 0x88f8, 0x6deb,
}
 
func errhandle(err error) {
  if err != nil {
    log.fatal(err)
  }
}
 
func load(name string) (m map[rune]int, err error) {
  f, err := os.open(name)
  if err != nil {
    return nil, err
  }
  defer f.close()
  buf := bufio.newreader(f)
  m = make(map[rune]int)
  var r rune
  for {
    r, _, err = buf.readrune()
    if err != nil {
      if err == io.eof {
        break
      }
      return nil, err
    }
    if r >= min_hanzi && r <= max_hanzi {
      m[r] += 1
    }
  }
  return m, nil
}
func classify(m map[rune]int) (idv []float64, dis float64) {
  len_m := len(m)
  for i, v := range labels {
    if debug {
      fmt.println(i, m[v], string(v), float64(m[v])/float64(len_m))
    }
    idv = append(idv, float64(m[v])/float64(len_m))
  }
  for _, v := range idv {
    dis += math.pow(v, 2)
  }
  dis = math.sqrt(dis)
  return
}
func check(fp string, dis float64) {
  switch {
  case dis >= limen:
    fmt.println(fp, dis, "涉黄")
  case dis == 1.0:
    fmt.println(fp, dis, "你在作弊吗")
  case dis == 0:
    fmt.println(fp, dis, "检查一下文件字符编码是不是utf8格式吧")
  default:
    fmt.println(fp, dis, "正常")
  }
}
 
func walkfunc(fp string, info os.fileinfo, err error) error {
  if path.ext(fp) == ".txt" {
    m, err := load(fp)
    errhandle(err)
    _, dis := classify(m)
    check(fp, dis)
  }
  return err
}
 
var file string
 
func init() {
  _, err := os.stat(data_dir)
  if err != nil {
    err = os.mkdir(data_dir, os.modeperm)
    errhandle(err)
  }
  flag.stringvar(&file, "file", "", "file read in,if you don't give the file read in,"+
    "it will create a data dictionary,just pust your files in it")
}
 
func main() {
  flag.parse()
  if file == "" {
    filepath.walk(data_dir, walkfunc)
    return
  }
  m, err := load(file)
  errhandle(err)
  _, dis := classify(m)
  check(file, dis)
 
}

以上所述就是本文的全部内容了，希望大家能够喜欢。