Entry Hsu:1992:CCP from ijcpol.bib

Last update: Sun Oct 15 02:20:53 MDT 2017                Valid HTML 3.2!

Index sections

Top | Math | A | B | C | D | E | F | G | H | I | J | K | L | M | N | O | P | Q | R | S | T | U | V | W | X | Y | Z

BibTeX entry

@Article{Hsu:1992:CCP,
  author =       "Loke Soo Hsu and Zhibiao Wu",
  title =        "{Chinese} character prediction by recurrent network",
  journal =      j-COMPUTER-PROC-CHINESE-ORIENTAL-LANG,
  volume =       "6",
  number =       "2",
  pages =        "179--194",
  month =        dec,
  year =         "1992",
  CODEN =        "CPCLE6",
  ISSN =         "0715-9048",
  bibdate =      "Mon Feb 25 07:56:32 2002",
  bibsource =    "ftp://ftp.math.utah.edu/pub/mirrors/ftp.ira.uka.de/bibliography/Ai/cogann.bib;
                 http://www.math.utah.edu/pub/tex/bib/ijcpol.bib",
  abstract =     "The speedy input of Chinese characters from a keyboard
                 is a problem that receives much attention. Many
                 different input methods have been implemented and put
                 into use. For all these input methods, the prediction
                 aids are provided. This requires the system to predict
                 the next most likely Chinese character based on the
                 previous input. The conventional approach is to use the
                 frequencies of the Chinese word phrases. The authors
                 present an approach by using the combination of the
                 simple recurrent neural network and the genetic
                 algorithm. The simple recurrent network makes the
                 prediction result change dynamically for different
                 contexts, even the previous Chinese character may be
                 the same. The genetic algorithm is used to do a
                 generalized adaptive search in the solution space for a
                 better input pattern encoding scheme to ensure that the
                 network size is not too large. Experiments show that
                 the method is promising.",
  fjournal =     "Computer Processing of Chinese and Oriental
                 Languages",
  journal-URL =  "http://www.worldscientific.com/worldscinet/ijcpol",
  keywords =     "genetic algorithms connectionism neural networks
                 cogann",
}

Related entries