<< 1 >>
Rating: 
Summary: Excellent
Review: I recommend this very highly. The words "online applications" in the title suggest that this book is about NLP for websites, but it's much more general than that; certainly any of the technologies discussed in it could in fact be implemented on a website, but "online applications" should be interpreted as meaning something like "applications that are made possible or commercially viable by the availability of large bodies of documents over the Internet."The focus of the book is on technologies with commercial applications, and that aspect of the topics discussed in the book is addressed clearly and well. However, there's also plenty in the book that will be of interest to a researcher, especially one looking for an overview of a topic. In fact, the book reads much like a series of well-written review articles. The first chapter of the book discusses some of the general issues and challenges in natural language processing, on a level
that should be accessible to pretty much anyone. (Actually, one of the really outstanding features of this book is its overall readability.) Subsequent chapters focus on information retrieval, information extraction, and text categorization. The final chapter has shorter sections on other topics, including summarization and named entity recognition. All chapters of the book are characterized by high readability, clear explanations of algorithms (and I say that as someone who struggles with ANY algorithm), and good explanations of the relevant evaluation metrics. This book would be a good starting point for anyone; if you're not a beginner in natural language processing, you'll still find much that's useful in this book.
Rating:
Summary: Excellent
Review: I recommend this very highly. The words "online applications" in the title suggest that this book is about NLP for websites, but it's much more general than that; certainly any of the technologies discussed in it could in fact be implemented on a website, but "online applications" should be interpreted as meaning something like "applications that are made possible or commercially viable by the availability of large bodies of documents over the Internet." The focus of the book is on technologies with commercial applications, and that aspect of the topics discussed in the book is addressed clearly and well. However, there's also plenty in the book that will be of interest to a researcher, especially one looking for an overview of a topic. In fact, the book reads much like a series of well-written review articles. The first chapter of the book discusses some of the general issues and challenges in natural language processing, on a level
that should be accessible to pretty much anyone. (Actually, one of the really outstanding features of this book is its overall readability.) Subsequent chapters focus on information retrieval, information extraction, and text categorization. The final chapter has shorter sections on other topics, including summarization and named entity recognition. All chapters of the book are characterized by high readability, clear explanations of algorithms (and I say that as someone who struggles with ANY algorithm), and good explanations of the relevant evaluation metrics. This book would be a good starting point for anyone; if you're not a beginner in natural language processing, you'll still find much that's useful in this book.
Rating: 
Summary: An excellent overview
Review: Work on Natural Language processing has been going on for at least
thirty years. In the past most natural language processing (NLP)
applications where mainly in the research realm. The rapid
increase in computer processing power and disk storage capacity
have moved NLP from research into the area of applied science.
This gives NLP the feel of a new and vibrant area. Progress is
being made rapidly, but the research literature can be difficult
for someone who has no experience with NLP. Simply learning the
terminology can be time consuming. This book by Jackson and Moulinier provides an excellent overview
of several sub-areas of NLP applied to natural language text.
Both Jackson and Moulinier have been involved in implementing
NLP applications in a commercial context, so there is a
some concentration on applying NLP in real applications, rather
than artificial contexts like the Message Understand Conference
data set. I purchased this book for its coverage of Information Extraction.
Along with this book I read a number of papers from the research
literature. One paper I found particularly interesting was
a paper on the FASTUS NLP system developed by researchers at
SRI. I was very happy to see that FASTUS and finite automata
approaches were covered in more detail in this book. For my purposes I would have liked a book that moved from
an overview of various NLP applications to more implementation
detail. For example, while I think that I understand
push down automata from working on parsers for compilers,
I don't fully understand them in the NLP context. This book
did not go into enough detail to make this clear. I cannot really offer this lack of detail as a criticism
since I don't believe that it was the authors intention to
provide this level of detail. Their objective is to provide
a detailed overview and I think that they succeeded in doing
this. A book that provided this overview with
details on implementation would be much longer, perhaps the
size of Manning and Schutze's excellent book "Foundations of
Natural Language Processing" (Manning and Schutze provide
a great deal of detail, but they do not cover information
extraction).
Rating:
Summary: An excellent overview
Review: Work on Natural Language processing has been going on for at least
thirty years. In the past most natural language processing (NLP)
applications where mainly in the research realm. The rapid
increase in computer processing power and disk storage capacity
have moved NLP from research into the area of applied science.
This gives NLP the feel of a new and vibrant area. Progress is
being made rapidly, but the research literature can be difficult
for someone who has no experience with NLP. Simply learning the
terminology can be time consuming. This book by Jackson and Moulinier provides an excellent overview
of several sub-areas of NLP applied to natural language text.
Both Jackson and Moulinier have been involved in implementing
NLP applications in a commercial context, so there is a
some concentration on applying NLP in real applications, rather
than artificial contexts like the Message Understand Conference
data set. I purchased this book for its coverage of Information Extraction.
Along with this book I read a number of papers from the research
literature. One paper I found particularly interesting was
a paper on the FASTUS NLP system developed by researchers at
SRI. I was very happy to see that FASTUS and finite automata
approaches were covered in more detail in this book. For my purposes I would have liked a book that moved from
an overview of various NLP applications to more implementation
detail. For example, while I think that I understand
push down automata from working on parsers for compilers,
I don't fully understand them in the NLP context. This book
did not go into enough detail to make this clear. I cannot really offer this lack of detail as a criticism
since I don't believe that it was the authors intention to
provide this level of detail. Their objective is to provide
a detailed overview and I think that they succeeded in doing
this. A book that provided this overview with
details on implementation would be much longer, perhaps the
size of Manning and Schutze's excellent book "Foundations of
Natural Language Processing" (Manning and Schutze provide
a great deal of detail, but they do not cover information
extraction).
<< 1 >>
| 
