Assignment 8 Final

.docx

School

University of Maryland, University College *

*We aren’t endorsed by this school

Course

661

Subject

Computer Science

Date

Jun 27, 2024

Type

docx

Pages

Uploaded by Janel0421

1 Assignment 8: Natural Language Processing (NLP) Janel Handy University of Maryland Global Campus HIMS 661: The Application of Information Technology in Healthcare Administration Dr. Craig Drayden July 14, 2023

2 Natural Language Processing (NLP) Natural language processing (NLP) is the ability of a computer program to understand human language as it is spoken and written which is referred to as natural language, a component of artificial intelligence (Burns, 2023). Therefore, NLP models are able to be utilized by finding relationships between letters, words, and sentences that are found in text datasets. NLP helps to streamline business operation and improve employee productivity. In order for applications to perform successfully, programmers have to develop tools that teach natural language driven applications to recognize and understand accurately from implementation. Natural Language Processing Tasks, Tools, & Approaches NLP tasks are helpful to assist with breaking down human text and voice data to help computers process the language. An example of tasks are speech recognition, name entity recognition, and natural language generation. Speech recognition which is referred to as speech txt has the ability of converting voice into text which is required for any application that follows voice commands or answers spoken questions (IBM, n.d.). Named entity recognition (NEM) uses words or phrases as useful entities. For example, they have the ability to recognize Maryland as a location. Another useful task is natural language generation which does speech to text and takes information into the human language. There are two useful tools that are utilized for NLP which are spaCy and Natural Language Toolkit (NLTK). NLTK programming language utilizes a vast number of tools for certain NLP tasks. Many of the tasks are found in the NLTK which is a useful resource of libraries, programs, and education resources for building NLP programs (IBM). SpaCy is considered one of the most versatile NLP approaches. This tool supports approximately 66

3 languages which uses pre-trained word vectors for named entity recognition, part-of-speech tagging, dependency parsing, sentence segmentation, text classification, lemmatization (Deep Learning AI). SpaCy is considered one of the more and efficient and faster tools which has become widely used. There are three type of NLP approaches which are rule-based approach, machine learning approach, and neural network approach. Rule-based approach is known as one of the oldest methods which helps to process textual data. The benefit of rule-based approach applies a particular set of rules or patterns to capture specific structures, extract information, or perform tasks such as text classifications. There are four steps to complete rule-based approach which are rule creation, rule application, rule processing, and rule refinement. In this approach, the rules are manually created and relies on linguistic or domain-specific knowledge but this approach can be challenging to handle complex language. A disadvantage of this approach is that it does not handle complex language applications. How to extract useful information using NLP The use of NLP plays a vital role of extracting diagnostic codes such as ICD-10-CM and codes CPT 4.0 from summary notes to reimburse claims. Staff follow a process when extracting information and the first step is data pre-processing. Data preprocessing is used to clean and standardize the text which ensures accuracy during the extraction process. Named entity recognition (NER) which involves identifying key information in the text and classification of a set of predefined categories. For progress notes, NER is useful for extracting medical terminology and information that is useful for coding (Deep Learning AI). Implementing rule based approaches are useful to help associate patters within progress notes that are corresponding with diagnostic codes. When the codes are successfully extracted techniques can be developed to

Your preview ends here

Eager to read complete document? Join bartleby learn and gain access to the full version