Submit Search
Upload
WordNet Based Online Reverse Dictionary with Improved Accuracy and Parts-of-Speech Sorting
•
0 likes
•
10 views
IRJET Journal
Follow
https://www.irjet.net/archives/V10/i11/IRJET-V10I1110.pdf
Read less
Read more
Engineering
Report
Share
Report
Share
1 of 6
Download now
Download to read offline
Recommended
Malayalam Word Sense Disambiguation using Machine Learning Approach
Malayalam Word Sense Disambiguation using Machine Learning Approach
IRJET Journal
IRJET-Semantic Similarity Between Sentences
IRJET-Semantic Similarity Between Sentences
IRJET Journal
Semantic Similarity Between Sentences
Semantic Similarity Between Sentences
IRJET Journal
LARQS: AN ANALOGICAL REASONING EVALUATION DATASET FOR LEGAL WORD EMBEDDING
LARQS: AN ANALOGICAL REASONING EVALUATION DATASET FOR LEGAL WORD EMBEDDING
kevig
LARQS: AN ANALOGICAL REASONING EVALUATION DATASET FOR LEGAL WORD EMBEDDING
LARQS: AN ANALOGICAL REASONING EVALUATION DATASET FOR LEGAL WORD EMBEDDING
kevig
IRJET- Short-Text Semantic Similarity using Glove Word Embedding
IRJET- Short-Text Semantic Similarity using Glove Word Embedding
IRJET Journal
THE ABILITY OF WORD EMBEDDINGS TO CAPTURE WORD SIMILARITIES
THE ABILITY OF WORD EMBEDDINGS TO CAPTURE WORD SIMILARITIES
kevig
THE ABILITY OF WORD EMBEDDINGS TO CAPTURE WORD SIMILARITIES
THE ABILITY OF WORD EMBEDDINGS TO CAPTURE WORD SIMILARITIES
kevig
Recommended
Malayalam Word Sense Disambiguation using Machine Learning Approach
Malayalam Word Sense Disambiguation using Machine Learning Approach
IRJET Journal
IRJET-Semantic Similarity Between Sentences
IRJET-Semantic Similarity Between Sentences
IRJET Journal
Semantic Similarity Between Sentences
Semantic Similarity Between Sentences
IRJET Journal
LARQS: AN ANALOGICAL REASONING EVALUATION DATASET FOR LEGAL WORD EMBEDDING
LARQS: AN ANALOGICAL REASONING EVALUATION DATASET FOR LEGAL WORD EMBEDDING
kevig
LARQS: AN ANALOGICAL REASONING EVALUATION DATASET FOR LEGAL WORD EMBEDDING
LARQS: AN ANALOGICAL REASONING EVALUATION DATASET FOR LEGAL WORD EMBEDDING
kevig
IRJET- Short-Text Semantic Similarity using Glove Word Embedding
IRJET- Short-Text Semantic Similarity using Glove Word Embedding
IRJET Journal
THE ABILITY OF WORD EMBEDDINGS TO CAPTURE WORD SIMILARITIES
THE ABILITY OF WORD EMBEDDINGS TO CAPTURE WORD SIMILARITIES
kevig
THE ABILITY OF WORD EMBEDDINGS TO CAPTURE WORD SIMILARITIES
THE ABILITY OF WORD EMBEDDINGS TO CAPTURE WORD SIMILARITIES
kevig
EasyChair-Preprint-7375.pdf
EasyChair-Preprint-7375.pdf
NohaGhoweil
IRJET- Semantic Question Matching
IRJET- Semantic Question Matching
IRJET Journal
IRJET- Automatic Text Summarization using Text Rank
IRJET- Automatic Text Summarization using Text Rank
IRJET Journal
Ijarcet vol-2-issue-7-2252-2257
Ijarcet vol-2-issue-7-2252-2257
Editor IJARCET
Ijarcet vol-2-issue-7-2252-2257
Ijarcet vol-2-issue-7-2252-2257
Editor IJARCET
Conceptual Similarity Measurement Algorithm For Domain Specific Ontology
Conceptual Similarity Measurement Algorithm For Domain Specific Ontology
Zac Darcy
G04124041046
G04124041046
IOSR-JEN
Conceptual similarity measurement algorithm for domain specific ontology[
Conceptual similarity measurement algorithm for domain specific ontology[
Zac Darcy
A018110108
A018110108
IOSR Journals
Co extracting opinion targets and opinion words from online reviews based on ...
Co extracting opinion targets and opinion words from online reviews based on ...
redpel dot com
Semantic web service discovery approaches
Semantic web service discovery approaches
IJCSES Journal
Context Sensitive Relatedness Measure of Word Pairs
Context Sensitive Relatedness Measure of Word Pairs
IJCSIS Research Publications
A SURVEY PAPER ON EXTRACTION OF OPINION WORD AND OPINION TARGET FROM ONLINE R...
A SURVEY PAPER ON EXTRACTION OF OPINION WORD AND OPINION TARGET FROM ONLINE R...
ijiert bestjournal
AN EMPIRICAL STUDY OF WORD SENSE DISAMBIGUATION
AN EMPIRICAL STUDY OF WORD SENSE DISAMBIGUATION
ijnlc
Wsd final paper
Wsd final paper
Milind Gokhale
CONTEXT BASED LEXICAL SIMPLIFICATION
CONTEXT BASED LEXICAL SIMPLIFICATION
IRJET Journal
Measure Term Similarity Using a Semantic Network Approach
Measure Term Similarity Using a Semantic Network Approach
BOHR International Journal of Intelligent Instrumentation and Computing
228-SE3001_2
228-SE3001_2
Boshra Albayaty
Effect of word embedding vector dimensionality on sentiment analysis through ...
Effect of word embedding vector dimensionality on sentiment analysis through ...
IAESIJAI
Analysis of Opinionated Text for Opinion Mining
Analysis of Opinionated Text for Opinion Mining
mlaij
TUNNELING IN HIMALAYAS WITH NATM METHOD: A SPECIAL REFERENCES TO SUNGAL TUNNE...
TUNNELING IN HIMALAYAS WITH NATM METHOD: A SPECIAL REFERENCES TO SUNGAL TUNNE...
IRJET Journal
STUDY THE EFFECT OF RESPONSE REDUCTION FACTOR ON RC FRAMED STRUCTURE
STUDY THE EFFECT OF RESPONSE REDUCTION FACTOR ON RC FRAMED STRUCTURE
IRJET Journal
More Related Content
Similar to WordNet Based Online Reverse Dictionary with Improved Accuracy and Parts-of-Speech Sorting
EasyChair-Preprint-7375.pdf
EasyChair-Preprint-7375.pdf
NohaGhoweil
IRJET- Semantic Question Matching
IRJET- Semantic Question Matching
IRJET Journal
IRJET- Automatic Text Summarization using Text Rank
IRJET- Automatic Text Summarization using Text Rank
IRJET Journal
Ijarcet vol-2-issue-7-2252-2257
Ijarcet vol-2-issue-7-2252-2257
Editor IJARCET
Ijarcet vol-2-issue-7-2252-2257
Ijarcet vol-2-issue-7-2252-2257
Editor IJARCET
Conceptual Similarity Measurement Algorithm For Domain Specific Ontology
Conceptual Similarity Measurement Algorithm For Domain Specific Ontology
Zac Darcy
G04124041046
G04124041046
IOSR-JEN
Conceptual similarity measurement algorithm for domain specific ontology[
Conceptual similarity measurement algorithm for domain specific ontology[
Zac Darcy
A018110108
A018110108
IOSR Journals
Co extracting opinion targets and opinion words from online reviews based on ...
Co extracting opinion targets and opinion words from online reviews based on ...
redpel dot com
Semantic web service discovery approaches
Semantic web service discovery approaches
IJCSES Journal
Context Sensitive Relatedness Measure of Word Pairs
Context Sensitive Relatedness Measure of Word Pairs
IJCSIS Research Publications
A SURVEY PAPER ON EXTRACTION OF OPINION WORD AND OPINION TARGET FROM ONLINE R...
A SURVEY PAPER ON EXTRACTION OF OPINION WORD AND OPINION TARGET FROM ONLINE R...
ijiert bestjournal
AN EMPIRICAL STUDY OF WORD SENSE DISAMBIGUATION
AN EMPIRICAL STUDY OF WORD SENSE DISAMBIGUATION
ijnlc
Wsd final paper
Wsd final paper
Milind Gokhale
CONTEXT BASED LEXICAL SIMPLIFICATION
CONTEXT BASED LEXICAL SIMPLIFICATION
IRJET Journal
Measure Term Similarity Using a Semantic Network Approach
Measure Term Similarity Using a Semantic Network Approach
BOHR International Journal of Intelligent Instrumentation and Computing
228-SE3001_2
228-SE3001_2
Boshra Albayaty
Effect of word embedding vector dimensionality on sentiment analysis through ...
Effect of word embedding vector dimensionality on sentiment analysis through ...
IAESIJAI
Analysis of Opinionated Text for Opinion Mining
Analysis of Opinionated Text for Opinion Mining
mlaij
Similar to WordNet Based Online Reverse Dictionary with Improved Accuracy and Parts-of-Speech Sorting
(20)
EasyChair-Preprint-7375.pdf
EasyChair-Preprint-7375.pdf
IRJET- Semantic Question Matching
IRJET- Semantic Question Matching
IRJET- Automatic Text Summarization using Text Rank
IRJET- Automatic Text Summarization using Text Rank
Ijarcet vol-2-issue-7-2252-2257
Ijarcet vol-2-issue-7-2252-2257
Ijarcet vol-2-issue-7-2252-2257
Ijarcet vol-2-issue-7-2252-2257
Conceptual Similarity Measurement Algorithm For Domain Specific Ontology
Conceptual Similarity Measurement Algorithm For Domain Specific Ontology
G04124041046
G04124041046
Conceptual similarity measurement algorithm for domain specific ontology[
Conceptual similarity measurement algorithm for domain specific ontology[
A018110108
A018110108
Co extracting opinion targets and opinion words from online reviews based on ...
Co extracting opinion targets and opinion words from online reviews based on ...
Semantic web service discovery approaches
Semantic web service discovery approaches
Context Sensitive Relatedness Measure of Word Pairs
Context Sensitive Relatedness Measure of Word Pairs
A SURVEY PAPER ON EXTRACTION OF OPINION WORD AND OPINION TARGET FROM ONLINE R...
A SURVEY PAPER ON EXTRACTION OF OPINION WORD AND OPINION TARGET FROM ONLINE R...
AN EMPIRICAL STUDY OF WORD SENSE DISAMBIGUATION
AN EMPIRICAL STUDY OF WORD SENSE DISAMBIGUATION
Wsd final paper
Wsd final paper
CONTEXT BASED LEXICAL SIMPLIFICATION
CONTEXT BASED LEXICAL SIMPLIFICATION
Measure Term Similarity Using a Semantic Network Approach
Measure Term Similarity Using a Semantic Network Approach
228-SE3001_2
228-SE3001_2
Effect of word embedding vector dimensionality on sentiment analysis through ...
Effect of word embedding vector dimensionality on sentiment analysis through ...
Analysis of Opinionated Text for Opinion Mining
Analysis of Opinionated Text for Opinion Mining
More from IRJET Journal
TUNNELING IN HIMALAYAS WITH NATM METHOD: A SPECIAL REFERENCES TO SUNGAL TUNNE...
TUNNELING IN HIMALAYAS WITH NATM METHOD: A SPECIAL REFERENCES TO SUNGAL TUNNE...
IRJET Journal
STUDY THE EFFECT OF RESPONSE REDUCTION FACTOR ON RC FRAMED STRUCTURE
STUDY THE EFFECT OF RESPONSE REDUCTION FACTOR ON RC FRAMED STRUCTURE
IRJET Journal
A COMPARATIVE ANALYSIS OF RCC ELEMENT OF SLAB WITH STARK STEEL (HYSD STEEL) A...
A COMPARATIVE ANALYSIS OF RCC ELEMENT OF SLAB WITH STARK STEEL (HYSD STEEL) A...
IRJET Journal
Effect of Camber and Angles of Attack on Airfoil Characteristics
Effect of Camber and Angles of Attack on Airfoil Characteristics
IRJET Journal
A Review on the Progress and Challenges of Aluminum-Based Metal Matrix Compos...
A Review on the Progress and Challenges of Aluminum-Based Metal Matrix Compos...
IRJET Journal
Dynamic Urban Transit Optimization: A Graph Neural Network Approach for Real-...
Dynamic Urban Transit Optimization: A Graph Neural Network Approach for Real-...
IRJET Journal
Structural Analysis and Design of Multi-Storey Symmetric and Asymmetric Shape...
Structural Analysis and Design of Multi-Storey Symmetric and Asymmetric Shape...
IRJET Journal
A Review of “Seismic Response of RC Structures Having Plan and Vertical Irreg...
A Review of “Seismic Response of RC Structures Having Plan and Vertical Irreg...
IRJET Journal
A REVIEW ON MACHINE LEARNING IN ADAS
A REVIEW ON MACHINE LEARNING IN ADAS
IRJET Journal
Long Term Trend Analysis of Precipitation and Temperature for Asosa district,...
Long Term Trend Analysis of Precipitation and Temperature for Asosa district,...
IRJET Journal
P.E.B. Framed Structure Design and Analysis Using STAAD Pro
P.E.B. Framed Structure Design and Analysis Using STAAD Pro
IRJET Journal
A Review on Innovative Fiber Integration for Enhanced Reinforcement of Concre...
A Review on Innovative Fiber Integration for Enhanced Reinforcement of Concre...
IRJET Journal
Survey Paper on Cloud-Based Secured Healthcare System
Survey Paper on Cloud-Based Secured Healthcare System
IRJET Journal
Review on studies and research on widening of existing concrete bridges
Review on studies and research on widening of existing concrete bridges
IRJET Journal
React based fullstack edtech web application
React based fullstack edtech web application
IRJET Journal
A Comprehensive Review of Integrating IoT and Blockchain Technologies in the ...
A Comprehensive Review of Integrating IoT and Blockchain Technologies in the ...
IRJET Journal
A REVIEW ON THE PERFORMANCE OF COCONUT FIBRE REINFORCED CONCRETE.
A REVIEW ON THE PERFORMANCE OF COCONUT FIBRE REINFORCED CONCRETE.
IRJET Journal
Optimizing Business Management Process Workflows: The Dynamic Influence of Mi...
Optimizing Business Management Process Workflows: The Dynamic Influence of Mi...
IRJET Journal
Multistoried and Multi Bay Steel Building Frame by using Seismic Design
Multistoried and Multi Bay Steel Building Frame by using Seismic Design
IRJET Journal
Cost Optimization of Construction Using Plastic Waste as a Sustainable Constr...
Cost Optimization of Construction Using Plastic Waste as a Sustainable Constr...
IRJET Journal
More from IRJET Journal
(20)
TUNNELING IN HIMALAYAS WITH NATM METHOD: A SPECIAL REFERENCES TO SUNGAL TUNNE...
TUNNELING IN HIMALAYAS WITH NATM METHOD: A SPECIAL REFERENCES TO SUNGAL TUNNE...
STUDY THE EFFECT OF RESPONSE REDUCTION FACTOR ON RC FRAMED STRUCTURE
STUDY THE EFFECT OF RESPONSE REDUCTION FACTOR ON RC FRAMED STRUCTURE
A COMPARATIVE ANALYSIS OF RCC ELEMENT OF SLAB WITH STARK STEEL (HYSD STEEL) A...
A COMPARATIVE ANALYSIS OF RCC ELEMENT OF SLAB WITH STARK STEEL (HYSD STEEL) A...
Effect of Camber and Angles of Attack on Airfoil Characteristics
Effect of Camber and Angles of Attack on Airfoil Characteristics
A Review on the Progress and Challenges of Aluminum-Based Metal Matrix Compos...
A Review on the Progress and Challenges of Aluminum-Based Metal Matrix Compos...
Dynamic Urban Transit Optimization: A Graph Neural Network Approach for Real-...
Dynamic Urban Transit Optimization: A Graph Neural Network Approach for Real-...
Structural Analysis and Design of Multi-Storey Symmetric and Asymmetric Shape...
Structural Analysis and Design of Multi-Storey Symmetric and Asymmetric Shape...
A Review of “Seismic Response of RC Structures Having Plan and Vertical Irreg...
A Review of “Seismic Response of RC Structures Having Plan and Vertical Irreg...
A REVIEW ON MACHINE LEARNING IN ADAS
A REVIEW ON MACHINE LEARNING IN ADAS
Long Term Trend Analysis of Precipitation and Temperature for Asosa district,...
Long Term Trend Analysis of Precipitation and Temperature for Asosa district,...
P.E.B. Framed Structure Design and Analysis Using STAAD Pro
P.E.B. Framed Structure Design and Analysis Using STAAD Pro
A Review on Innovative Fiber Integration for Enhanced Reinforcement of Concre...
A Review on Innovative Fiber Integration for Enhanced Reinforcement of Concre...
Survey Paper on Cloud-Based Secured Healthcare System
Survey Paper on Cloud-Based Secured Healthcare System
Review on studies and research on widening of existing concrete bridges
Review on studies and research on widening of existing concrete bridges
React based fullstack edtech web application
React based fullstack edtech web application
A Comprehensive Review of Integrating IoT and Blockchain Technologies in the ...
A Comprehensive Review of Integrating IoT and Blockchain Technologies in the ...
A REVIEW ON THE PERFORMANCE OF COCONUT FIBRE REINFORCED CONCRETE.
A REVIEW ON THE PERFORMANCE OF COCONUT FIBRE REINFORCED CONCRETE.
Optimizing Business Management Process Workflows: The Dynamic Influence of Mi...
Optimizing Business Management Process Workflows: The Dynamic Influence of Mi...
Multistoried and Multi Bay Steel Building Frame by using Seismic Design
Multistoried and Multi Bay Steel Building Frame by using Seismic Design
Cost Optimization of Construction Using Plastic Waste as a Sustainable Constr...
Cost Optimization of Construction Using Plastic Waste as a Sustainable Constr...
Recently uploaded
Current Transformer Drawing and GTP for MSETCL
Current Transformer Drawing and GTP for MSETCL
DeelipZope
main PPT.pptx of girls hostel security using rfid
main PPT.pptx of girls hostel security using rfid
NikhilNagaraju
power system scada applications and uses
power system scada applications and uses
DevarapalliHaritha
CCS355 Neural Network & Deep Learning Unit II Notes with Question bank .pdf
CCS355 Neural Network & Deep Learning Unit II Notes with Question bank .pdf
Asst.prof M.Gokilavani
🔝9953056974🔝!!-YOUNG call girls in Rajendra Nagar Escort rvice Shot 2000 nigh...
🔝9953056974🔝!!-YOUNG call girls in Rajendra Nagar Escort rvice Shot 2000 nigh...
9953056974 Low Rate Call Girls In Saket, Delhi NCR
9953056974 Call Girls In South Ex, Escorts (Delhi) NCR.pdf
9953056974 Call Girls In South Ex, Escorts (Delhi) NCR.pdf
9953056974 Low Rate Call Girls In Saket, Delhi NCR
VIP Call Girls Service Hitech City Hyderabad Call +91-8250192130
VIP Call Girls Service Hitech City Hyderabad Call +91-8250192130
Suhani Kapoor
HARMONY IN THE NATURE AND EXISTENCE - Unit-IV
HARMONY IN THE NATURE AND EXISTENCE - Unit-IV
RajaP95
Call Girls Delhi {Jodhpur} 9711199012 high profile service
Call Girls Delhi {Jodhpur} 9711199012 high profile service
rehmti665
Software and Systems Engineering Standards: Verification and Validation of Sy...
Software and Systems Engineering Standards: Verification and Validation of Sy...
VICTOR MAESTRE RAMIREZ
Biology for Computer Engineers Course Handout.pptx
Biology for Computer Engineers Course Handout.pptx
DeepakSakkari2
Study on Air-Water & Water-Water Heat Exchange in a Finned Tube Exchanger
Study on Air-Water & Water-Water Heat Exchange in a Finned Tube Exchanger
Anamika Sarkar
High Profile Call Girls Nagpur Meera Call 7001035870 Meet With Nagpur Escorts
High Profile Call Girls Nagpur Meera Call 7001035870 Meet With Nagpur Escorts
Call Girls in Nagpur High Profile
What are the advantages and disadvantages of membrane structures.pptx
What are the advantages and disadvantages of membrane structures.pptx
wendy cai
young call girls in Rajiv Chowk🔝 9953056974 🔝 Delhi escort Service
young call girls in Rajiv Chowk🔝 9953056974 🔝 Delhi escort Service
9953056974 Low Rate Call Girls In Saket, Delhi NCR
chaitra-1.pptx fake news detection using machine learning
chaitra-1.pptx fake news detection using machine learning
misbanausheenparvam
VICTOR MAESTRE RAMIREZ - Planetary Defender on NASA's Double Asteroid Redirec...
VICTOR MAESTRE RAMIREZ - Planetary Defender on NASA's Double Asteroid Redirec...
VICTOR MAESTRE RAMIREZ
Gfe Mayur Vihar Call Girls Service WhatsApp -> 9999965857 Available 24x7 ^ De...
Gfe Mayur Vihar Call Girls Service WhatsApp -> 9999965857 Available 24x7 ^ De...
srsj9000
Gurgaon ✡️9711147426✨Call In girls Gurgaon Sector 51 escort service
Gurgaon ✡️9711147426✨Call In girls Gurgaon Sector 51 escort service
jennyeacort
Call Us -/9953056974- Call Girls In Vikaspuri-/- Delhi NCR
Call Us -/9953056974- Call Girls In Vikaspuri-/- Delhi NCR
9953056974 Low Rate Call Girls In Saket, Delhi NCR
Recently uploaded
(20)
Current Transformer Drawing and GTP for MSETCL
Current Transformer Drawing and GTP for MSETCL
main PPT.pptx of girls hostel security using rfid
main PPT.pptx of girls hostel security using rfid
power system scada applications and uses
power system scada applications and uses
CCS355 Neural Network & Deep Learning Unit II Notes with Question bank .pdf
CCS355 Neural Network & Deep Learning Unit II Notes with Question bank .pdf
🔝9953056974🔝!!-YOUNG call girls in Rajendra Nagar Escort rvice Shot 2000 nigh...
🔝9953056974🔝!!-YOUNG call girls in Rajendra Nagar Escort rvice Shot 2000 nigh...
9953056974 Call Girls In South Ex, Escorts (Delhi) NCR.pdf
9953056974 Call Girls In South Ex, Escorts (Delhi) NCR.pdf
VIP Call Girls Service Hitech City Hyderabad Call +91-8250192130
VIP Call Girls Service Hitech City Hyderabad Call +91-8250192130
HARMONY IN THE NATURE AND EXISTENCE - Unit-IV
HARMONY IN THE NATURE AND EXISTENCE - Unit-IV
Call Girls Delhi {Jodhpur} 9711199012 high profile service
Call Girls Delhi {Jodhpur} 9711199012 high profile service
Software and Systems Engineering Standards: Verification and Validation of Sy...
Software and Systems Engineering Standards: Verification and Validation of Sy...
Biology for Computer Engineers Course Handout.pptx
Biology for Computer Engineers Course Handout.pptx
Study on Air-Water & Water-Water Heat Exchange in a Finned Tube Exchanger
Study on Air-Water & Water-Water Heat Exchange in a Finned Tube Exchanger
High Profile Call Girls Nagpur Meera Call 7001035870 Meet With Nagpur Escorts
High Profile Call Girls Nagpur Meera Call 7001035870 Meet With Nagpur Escorts
What are the advantages and disadvantages of membrane structures.pptx
What are the advantages and disadvantages of membrane structures.pptx
young call girls in Rajiv Chowk🔝 9953056974 🔝 Delhi escort Service
young call girls in Rajiv Chowk🔝 9953056974 🔝 Delhi escort Service
chaitra-1.pptx fake news detection using machine learning
chaitra-1.pptx fake news detection using machine learning
VICTOR MAESTRE RAMIREZ - Planetary Defender on NASA's Double Asteroid Redirec...
VICTOR MAESTRE RAMIREZ - Planetary Defender on NASA's Double Asteroid Redirec...
Gfe Mayur Vihar Call Girls Service WhatsApp -> 9999965857 Available 24x7 ^ De...
Gfe Mayur Vihar Call Girls Service WhatsApp -> 9999965857 Available 24x7 ^ De...
Gurgaon ✡️9711147426✨Call In girls Gurgaon Sector 51 escort service
Gurgaon ✡️9711147426✨Call In girls Gurgaon Sector 51 escort service
Call Us -/9953056974- Call Girls In Vikaspuri-/- Delhi NCR
Call Us -/9953056974- Call Girls In Vikaspuri-/- Delhi NCR
WordNet Based Online Reverse Dictionary with Improved Accuracy and Parts-of-Speech Sorting
1.
International Research Journal
of Engineering and Technology (IRJET) e-ISSN: 2395-0056 Volume: 10 Issue: 11 | Nov 2023 www.irjet.net p-ISSN: 2395-0072 © 2023, IRJET | Impact Factor value: 8.226 | ISO 9001:2008 Certified Journal | Page 81 WordNet Based Online Reverse Dictionary with Improved Accuracy and Parts-of-Speech Sorting Asha Chandran T1, Sajina K2 1 Lecturer In Computer Engineering, Department of Computer Engineering, Govt. Womens’ Polytechnic College Kayamkulam, Kerala, India 2 Lecturer In Computer Engineering, Department of Computer Engineering, Govt. Polytechnic College Neyyattinkara, Kerala, India ---------------------------------------------------------------------***--------------------------------------------------------------------- Abstract - Reverse dictionaries are the ones which map a phrase or short sentence to one ormorewordswhosemeaning matches these phrases. The input to a reverse dictionary problem is a regular forward dictionary itself. But simply mapping each meaning phrase in a forward dictionary does not serve the complete requirement for a reverse dictionary. This is because the user input phrase may not be exactly similar to that one in a forward dictionary. The input phrase may contain synonyms of the words in the meaning phrase. So the mapping problem aims to find a set of words in the forward dictionary which contains words or synonyms of the words in the meaning phrase. The reverse dictionary is particularly useful for writers, journalists, dataminingexperts and even to the general public to find a suitable single word for a phrase in their ideas. The paper aims to refine the RD output to include most similar words and eliminate unrelated words. Key Words: Synset, Ranking , Forward MappingSets(FMS), Reverse Mapping Set (RMS), Parts-of-Speech, WordNet , Query 1.INTRODUCTION Effective writing is an important concerntomanycategories of people like professional writers, students, scientists, teachers, marketing and advertisement professionals etc. Clarity and brevity are two important factors of effective writing. Brevity means being precise on the topic and clarity means using the apt words instead of beating the bushes. Usually people find it difficulttochoosethesuitablewordsin particular contexts. In fact, for most people with a certain level of education, the problem is often not lacking knowledge of the meaning of a word, but, rather, being unable to recall the appropriate word on demand. Here comes the role of a reverse dictionary (RD). Effectively, the RD addresses the “word is on the tip of my tongue, butIcan’t quite remember it” problem. A forward dictionary contains a set of words and each word has a set of definition phrases. To identify a single word equivalent to a phrase entered by the user, we need to compare the phrase with each definition in the dictionary, i.e., we need to find the semantic similarity betweentheuser input phrase and dictionary definitions. Semantic similarity has gained great importance over long time in many fields such as artificial intelligence, natural language processing, data mining etc. By addressing the RD creation problem, we develop better methods for semantic similarity identification. These methods can be used for a wide range of applications such as question-answersystems,plagiarismdetection, meta data mining,documentcompression,documentclassification etc. A regular (forward) dictionary maps words to their definitions, whereas a RD performs the converse mapping, i.e., given a phrase describing the desired concept, it provides words whose definitions match the entered definition phrase.Forexample,supposea forwarddictionary informs the user that the meaning of the word “spelunking” is “exploring caves.” A reverse dictionary, on the other hand, offers the user an opportunity to enter the phrase “check out natural caves” as input, and expect to receive the word “spelunking” (and possibly other words with similar meanings) as output. To construct a forward dictionary, we need a forward dictionary at our disposal. Upon receiving a search concept, the RD consults the forward dictionary and selects those words whose definitions are similar to this concept. These words then form the output of this RD lookup. The problem reduces to a concept similarity problem (CSP) .The CSP is a well-known hard problem which has been addressed in a number of ways with a limited degree of success. The RD problem can be thought of as a real-time online concept similarity problem. But there are many key differencesbetweenRDproblemand CSP which make direct use of existing results infeasible. 1.1 Related Works Most of the related works fall into the concept similarity identification attempts. In concept similarity identification, the domain may or may not be predefined. In the case of domain specificapplications, systems may have ahigherrate of accuracy since the list of words/concepts in specific domains will be more or less fixed. The concepts, in such
2.
International Research Journal
of Engineering and Technology (IRJET) e-ISSN: 2395-0056 Volume: 10 Issue: 11 | Nov 2023 www.irjet.net p-ISSN: 2395-0072 © 2023, IRJET | Impact Factor value: 8.226 | ISO 9001:2008 Certified Journal | Page 82 cases, will be easily distinguishable and hencelesseffortmay be needed for sense disambiguation. For the same reasons, domain specificontologiesmayconvergefastertooutput.But in RD creation, the inputcomesfromrealworldinstancesand similarity checking tends to be generic in nature and hence no specific restrictions are placed in domains. Concept vector creation and similarity calculations based on cosine of vectors[3] is a domainspecificapproachtomeasure similarity between concepts. Here words are considered as concepts. An m-nodehierarchy in a corpus can be mappedto an m-nodeconcept hierarchywitheachnodeinthehierarchy having an m-dimensional concept vector. The method, while being suitable fordomain specific applications,stillbearsthe load of constructing high dimensional vectors. Also high dimensionality of vectors cause thesentencesrepresentedas sparse. Latent Semantic Indexing [6] is a mathematical technique used to reduce the dimensionality of concept vectors without sacrificing their quality. LSI is thus a dimensionality reduction technique. In the case of multiword similarity, the works are usually concentrated on paragraphs matching. The paragraphs are considered as short documents itself. In such cases, the system has sufficient contextual information to compare [8]. Hence these methods cannot be directly applied in RD creation. Similarity identification between phrases or short sentences can be done using bipartite matching [2]. This work is more relevant in RD construction since inputs to RD also lie in the category of ‘short sentences’. The system assumes words in each sentence to be matched as nodes in a bipartite graph. The edges between nodes are weightedusinganyoftheword similarity computation methods. Overall similarity between sentences is measured by determining maximum bipartite matching between the two sets. Another approach identifiessimilarity betweensentencesby constructing a word order vector and similarity vector[5]. A word order vector is constructed from the joint word set of the sentences to be compared. A raw semantic vector is constructed from the lexical database and the joint word set. A semantic vector is constructed from the raw semantic vector and the corpus. Semantic similarity is calculated from the semantic vectors of the sentences. Similarly, order similarity is calculated from the order vectors of the sentences. Overall similarity is calculated from the semantic similarity and the word order similarity. Many other works try to identify similarity between sentences or phrase by calculating similarity between their constituent words. Similarity identification methods are broadly classified into corpus based and knowledge based. A number of such methods are described in [4]. 1.2 Existing Reverse Dictionary techniques Wordster Reverse Dictionary based on semantic relationships The reverse dictionary system under our considerationis Wordster Reverse Dictionary (WRD) [1]. The systemtakesa phrase entered by the user as input and constructs a query based on it. The query is a Boolean expression containing words other than stop words connected using the logical operand AND. A. Steps in RD Execution a. RD Creation. b. Input query generation from user input phrase. c. RD querying (Query Execution) d. Probable Query Expansion (if sufficient results are not obtained). e. Ranking of candidate output words f. Sorting of candidate words based on ranking RD Creation To start with, we need to create the reverse mapping of a forward dictionary. We construct reverse mapping set for the forward dictionary as follows. For every word w in the forward dictionary, the algorithm takes each sense phrase (meaning phrase) of the word. Each word in the sense phrase is stemmed to its root form. The popular Porter Stemming algorithm [9] is used for this. The word w is then included in the reverse mapping set (RMS) of each of the stemmed word. RD Querying Next, we need to query the RMS to find the possible output words (candidate words) for an input phrase. For this, the user input phrase U has to be modified to the form of a query. The GenerateQuery algorithm does this as follows. From the user input phrase U, the level 1stopwords are removed. The remaining terms are connectedusingAND to form a Boolean expression Q. Q is again modified by expanding antonyms present in it. The Expand Antonyms algorithm takes a query as input and if any negated terms such as not, nor, neither etc. are present in the query, antonyms of the succeeding word is retrieved. The negated term ti is replaced with its antonyms connected using OR. The new query is termed as Q’. The GenerateQuery algorithm takes the output Q’andcombineit to Q using OR. Each term ti in Q is stemmed to get ti^ and in Q, it gets replaced as ( ti OR ti^ ). The terms in Q is reordered such that all nouns appear before verbs and all verbs appear before adjectives and adverbs. By thisstep,the user input phrase is converted to the RMS query Q.
3.
International Research Journal
of Engineering and Technology (IRJET) e-ISSN: 2395-0056 Volume: 10 Issue: 11 | Nov 2023 www.irjet.net p-ISSN: 2395-0072 © 2023, IRJET | Impact Factor value: 8.226 | ISO 9001:2008 Certified Journal | Page 83 The query can now be executed. The ExecuteQuery algorithm simply takes set intersection of the RMS of the constituent terms. Since the RMS now contains only integer values corresponding to the word id s, this is a fast arithmetic operation. This algorithm returns a setofwordid s of the candidate output words. In case if the ExecuteQuery algorithm does not return sufficient number of candidate words(α) as specified in the application, the query is expanded using other conceptually similar terms such as synonyms, hyponyms, hypernyms etc. The ExpandQuery algorithm takes each term ti in the input query Q and replace ti with ( ti OR syn(ti ) ) or ( ti OR hypo(ti ) ) or ( ti OR hyper(ti ) ) according to the argument type being synonym, hyponym,hypernyms respectively.The expansion process is done sequentially starting from synonyms set. After expanding using synonyms, the algorithm checks whether the required no. of results are obtained. If yes, they are sorted. Otherwise, expansion is done using the second type, i.e, hyponyms,andtheprocessis repeated. Finally expansion using hypernyms are done. If none of these expansions help in retrieving required number of terms, then the GenerateQuery algorithm computes the cardinality of RMS of each term ti . The terms are then sorted in the decreasing orderofthecardinalityand terms with highest cardinality are removed from the query. Again, the algorithm checks whether the required no α is reached. If not, the above process is repeated until the number of terms is reduced to two. Once α output words are obtained, they are sorted. The reason for removing terms with higher cardinality is that they reduce the probability to get enough word id s while taking intersections. The removal is done until there only two terms, because it needs at least two sets to perform an intersection or union. Ranking and Sorting the Candidate words The output words are to besortedonthedecreasingorder of their similarity with the input phrase. The SortResults algorithm achieves this by considering two factors of each output words- term similarity and term importance. Term similarity ρ is computed between every pair of terms (a,b) where a ε S and b ε U. Term importance λ(a,S) indicate how critical the term a is in the context of the phrase S. Also, λ(b,U) is calculated. Similarity measure between S and U, µ(a,S,b,U), is calculated as the product of these three terms ρ(a,b), λ(a,S) and λ(b,U). Architecture for Wordster Reverse Dictionary Fig-1 Dictionary Architecture The input query is passed to the RDA module which take each valid term in the query and access the database for the term’s RMS, synonyms, hyponyms and hypernyms. The database is constructed from WordNet 3.1 version. A pool of threads accomplish the simultaneous access of RMS and relationship sets. This makes a faster and parallel execution of the query possible. A cache is used to store these databases. According to the size of the database,cachesizeis designed to accommodate these sets. OneLook.com and Dictionary.com OneLook.com [11] and Dictionary.com[7]aretworeverse dictionary systems which are alreadyavailableonline.These systems provide up to a maximum of 100 output words matching to the user input phrase. But the quality of the solution of the Wordster approach is better than these from a set of experimental results. 2. PROPOSED REVERSE DICTIONARY SYSTEM Consider an input phrase, ‘study of sound’. The expected result is ‘acoustics’ .When evaluating the direct expression, we get the candidate words ‘acoustics, ’echogram’ and ‘echography’. These results are perfectly acceptable since they all lie in the category of study about sound. Since we got only three candidate words and the user requires at least ten words, we go to the indirect expression. Indirect expressions are formed using synonyms, hyponyms and hypernyms of the input words. Now let us see the candidate words from the result of indirect expression evaluation. The indirect expressions are constructed as follows: (Study OR synonym(study)) AND ((sound OR synonym(sound)) If sufficient results are not obtained, the hyponyms are used for query. (Study OR hyponym(study)) AND ((sound OR hyponym(sound)) If still wedon’t obtain sufficient results, hypernyms aretried. ((Study OR hypernym(study)) AND ((sound OR hypernym(sound))
4.
International Research Journal
of Engineering and Technology (IRJET) e-ISSN: 2395-0056 Volume: 10 Issue: 11 | Nov 2023 www.irjet.net p-ISSN: 2395-0072 © 2023, IRJET | Impact Factor value: 8.226 | ISO 9001:2008 Certified Journal | Page 84 Accordingly we get many indirect expressions since a word can have many synonyms, hyponyms and hypernyms. Surprisingly, the only one result from one of the indirect expressions is the word ‘Aquinas’ which is the name of an ancient Italian theologian!!! 2.1 Reason for Unrelated Output words We can see that even after several similarity checking and sorting based on relevance, the output contain certain erroneous terms. This may arise because we are trying to obtain set intersections of the RMS s of the terms in the user query. If in case a more appropriate word appears in the intersection of two of the synset id sets out of three sets, for example, but not in the third set, that candidate word is eliminated in the existing algorithm. When a query is executed, if sufficient numbers of candidate words are not obtained, the query is expanded to include synonyms, hyponyms and hypernyms of the constituent terms respectively. While including hypernyms of a term,we are traversing up in the WordNethierarchy.Goinghigher,the terms tend to be more general in nature. This may result in a higher probability for obtaining results in set intersection, since these terms, being general in nature, will appear in the dictionary definitions of many words. Thiscause a more appropriate word to bediscardedfromthe list and more general or unrelated words to be added in the list. Also, when the user enters a phrase, if we can determine which Parts-of-Speech is used in the input phrase, we can sort the candidate words according to this Parts-of-Speech. This will help in reducing user frustration when he/she expects a noun and the returned output is an adjective or verb. Our efforts are pointed towards eliminating these problems. 2.2 Methods to Eliminate Unrelated words from Output Elimination of anappropriatewordcanbeavoidedbyslightly changing the ranking process so as to accommodate issue inherent in the set intersection problem. The ranking has to be modified such that acandidate word that appearsinmout of n RMS sets of the words in the input query will be assigned an appropriate rank. 2.3 New Ranking Algorithm Inputs: User input phrase with stop words removed. Output: A set of candidatewordswhosedictionarydefinition matches the user input phrase. Database: words, RMS, synsets, synonyms, hyponyms and hypernyms from WordNet 3.1 Let the user input phrase U consists of n words after removing stop words. Let the word IDs corresponding to the n words be w1,w2,…wn. RMS s of each of these terms are named as S1,S2,…Sn. Each of these sets will be a set of word IDs t1,t2,…tm. Let the user requires α candidate words. Algorithm: 1. Execute the direct expression, i.e., take set intersection of RMS of each term in the input query. 2. Check if the required numbers of candidate words are obtained. If not goto step 3 3. For i=1 to n, do steps 4 and 5. 4. For each ti ε Si, do the following step. 5. If ti is present in Sj, assign rank(ti)=rank(ti)+1/n. [rank(ti) is initialized to zero.] 6. Select all ti ε Si with rank(ti)> preset threshold. Let the new set be C with cardinality m . 7. For i=1 to m, do the following. 8. For j=1 to n, do the following. 9. Check if ti ε Sj. If not, take the word id wj corresponding to the set Sj. 10. Replace Sj with synonym set of wj and Check if ti ε Sj. 11. If yes, set rank(ti)=rank(ti)+(1/2)n Elsegoto next step. 12. Replace Sj with hyponym set and Check if ti ε Sj. 13. If yes, set rank(ti)=rank(ti)+(1/3)n 14. Sort the terms based on descending order of their rank. 15. Return the candidate words of the terms. 2.4 Pseudocode 1. Construct the query Q from the user input phrase U . Q=t1 AND t2 AND …AND tn 2. Execute the direct expression, i.e., take set intersection of RMS of each term in the input query. O=RMS(t1) ˄ RMS(t2) ˄ …. ˄ RMS(tn) 3. If |O|>= α goto step 14. Else goto next step. 4. For i=1 to n For j=1 to m If tj ε Si Set rank(tj)=rank(tj)+1/n. [rank(tj) is initialized to zero.] 5. For each Si in the set [ i:1…n] For j=1 to m If rank(tj)> preset threshold and tj not in O Add tj to output word set O. Else add tj to candidate word list C. 6. If |O|>= α goto step 14 . Else goto next step. 8. For i=1 to |C|
5.
International Research Journal
of Engineering and Technology (IRJET) e-ISSN: 2395-0056 Volume: 10 Issue: 11 | Nov 2023 www.irjet.net p-ISSN: 2395-0072 © 2023, IRJET | Impact Factor value: 8.226 | ISO 9001:2008 Certified Journal | Page 85 For j=1 to n If ti not in Sj take the word id wj corresponding to the set Sj. Replace Sj with synonym set of wj. If ti ε Syn(wj) { rank(ti)=rank(ti)+(1/2)n goto step 11 } Else { Replace Sj with hyponym set of wj. If ti ε Hypo(wj) { rank(ti)=rank(ti)+(1/3)n goto step 11 } } Else { Replace Sj with hypernym set of wj. If ti ε Hyper(wj) Rank(ti)=rank(ti)+(1/4)n } 9. For i=1 to |C| If rank(ti)> preset threshold Add ti to O. 10. Sort the terms in O based on descending order of their rank. 11. Return the sorted list of candidate words . The algorithm searches the semantic relationships for only those words whose RMS does not give any results for set intersection with the RMS sets of other words. Words in the input query whoseRMScontaincommoncandidatewordIDs are not checked further for synonym, hyponym and hypernyms relations. Hence unnecessary reduction of accuracy is avoided. 2.5 Identifying Parts-of-Speech from User Input Phrase WordNet contains nearly1,55,300 words. Out of these nearly 1,17,800 words are nouns. Of the remaining, 11500 words are verbs , 21200 are adjectives and 4500 are adverbs approximately. Since nouns contribute most of the vocabulary, there is a higher probability that the expected word to be a noun. We assume that if at least half of the valid words in the query are nouns, the expected parts-of-speech is a noun. In case an adjective is expected, the query will contain either adjective words or adjective indicators like ‘being’, ‘having ’,’like’ etc. Similarly, we can identify whether user expects a verb or adverb according to the parts-of-speech in which most of the words in the query fall in. Once the expected parts-of-speech is identified, the list of candidate words obtained as the result of our new ranking algorithm can be further refined such that those words with the expected parts-of-speech come first in the list. This may help to provide the user with an output that matches his/her requirements as far as possible. 3. CONCLUSIONS The paper describes how a reverse dictionary can be built, used and improved to match the user requirements. The paper proposes a new ranking algorithm that eliminates generic words and unrelated words from the candidate output word list. Also, it ensuresthataneligiblesimilarword will be added to the candidate words list in the appropriate position. Identification of parts-of-speech isdone byreading between the lines. This also improves the quality of the output. REFERENCES [1] Ryan Shaw, Debra Vander Meer and Kaushik Dutta ”Building a Scalable Database- Driven Reverse Dictionary”, vol. 25,no. 3,march 2013. [2] T. Dao and T. Simpson, “Measuring Similarity between Sentences, ”http://opensvn.csie.org/WordNetDotNet/trunk/Projects/T hanh/Paper/WordNetDotNet_Semantic_Similarity.pdf Oct. 2009), 2009 [3] J. Kim and K. Candan, “Cp/cv: Concept Similarity Mining without Frequency Information from Domain Describing Taxonomies,”Proc. ACM Conf. Information and Knowledge Management,2006. [4] R. Mihalcea, C. Corley, and C. Strapparava, “Corpus- Based and Knowledge-Based Measures of Text Semantic Similarity,” Proc.Nat’l Conf. Artificial Intelligence, 2006 [5] A Yuhua Li, David McLean, Zuhair A. Bandar, James D. O’Shea, and Keeley Crockett ” Sentence Similarity Based on Semantic Nets and Corpus Statistics”, IEEE TRANSACTIONS ON KNOWLEDGE AND DATA ENGINEERING, VOL. 18, NO. 8, AUGUST 2006 [6] Barbara Rosario,”Latent Semantic Indexing: An overview” Infosys 240 Spring 2000 Final Paper [7] Dictionary.com,LLC,“Reverse Dictionary,”http://dictionary.reference.com/reverse,2009. [8] E.Gabrilovich and S.Markovitch, “Wikipedia-Based Semantic Interpretation for Natural Language Processing,” J. Artificial Intelligence Research, vol. 34, no. 1, pp. 443-498, 2009.
6.
International Research Journal
of Engineering and Technology (IRJET) e-ISSN: 2395-0056 Volume: 10 Issue: 11 | Nov 2023 www.irjet.net p-ISSN: 2395-0072 © 2023, IRJET | Impact Factor value: 8.226 | ISO 9001:2008 Certified Journal | Page 86 [9] MPorter, “The Porter Stemming Algorithm,”http://tartarus. org/martin/PorterStemmer/, 2009. [10] O.S. Project “Opennlp,” http://opennlp.sourceforge.net/, 2009. [11] OneLook.com,“Onelook.comReverseDictionary,” [12] U. of Pennsylvania, “The Penn Treebank Project,” http://www. cis.upenn.edu/ treebank/, 2009.w.onelook.com/, 2009.
Download now