Phonetic Matching Research Articles

SoundexGR: An algorithm for phonetic matching for the Greek language

AbstractText usually suffers from typos which can negatively affect various Information Retrieval and Natural Language Processing tasks. Although there is a wide variety of choices for tackling this issue in the English language, this is not the case for other languages. For the Greek language, most of the existing phonetic algorithms provide rather insufficient support. For this reason, in this paper, we introduce an algorithm for phonetic matching designed for the Greek language: we start from the original Soundex and we redesign and extend it for accommodating the Greek language’s phonetic rules, ending up to a family of algorithms, that we call ${\tt Soundex}_{GR}$ . Then, we report various experimental results showcasing how the algorithm behaves in different scenarios, and we provide comparative results for various parameters of the algorithm for revealing the trade-off between precision and recall in datasets with different kinds of errors. We also provide comparative results with matching using stemming, full phonemic transcription, and edit distance, that demonstrate that ${\tt Soundex}_{GR}$ performs better (indicatively, it achieves F-Score over 95% in collections of similar-sounded words). The simplicity, efficiency, and effectiveness of the proposed algorithm make it applicable and adaptable to a wide range of tasks.

Creating A Validation Dataset to Test Record Linkage Algorithms in A Province-Wide South African Health Information Exchange

IntroductionIncreasing use of digital medical records creates disparate data resources for the same health care client population; and harnessing the benefits of real-time health data requires effective data linkage. A South African Health Information Exchange (HIE) collates and links routine health data from multiple sources, running daily updates through an automated ETL process. Many existing deterministic and probabilistic algorithms link person-level data using demographic identifiers, and can be combined in an optimised methodological pipeline. The performance of such pipelines must be validated against known matched pairs. The HIE uses current algorithms for record linkage, but methods that rely on similar spelling, name frequency and phonetic matching have been optimised for non-African names, and are not as effective.
 ObjectivesWe assessed common problems arising in the linkage process in the HIE, using this information to compile a curated representative African validation database for optimising existing and new linkage pipelines.
 ResultsUsing current linkage algorithms, we have identified the proportion of duplicates in the last five years, ranging from 25% in 2015 and stabilising at 10% by 2019. Common causes of duplicates across the whole database include mismatch in first name (37%), surname (17%), date of birth (13%), sex (8%) and South African Identification Number (0.2%). Complications from new-born naming and records of twins affect >8% of all records, and temporary health identifiers assigned at birth, during emergency response, and during poor connectivity of facilities to the provincial patient master index affect 2% of records.
 ConclusionsBased on these data, we have constructed a South African-specific, representative validation dataset that contains linkage pairs that represent placeholder phrases for newborns prior to naming (e.g. “baby of”), language variations; twins; character insertions, substitution and omissions in names with similar spellings; frequencies of names in the general population; and similar-sounding names.

Translation Challenges in Rendering Idiolects of Literary Characters

The article focuses on the challenges in rendering idiolects of literary characters. Idiolect as a means of speech characterisation of personages enables the researchers to see personages as linguistic personalities. Idiolects can fulfil several functions: comparative, psychological, distinctive, and characterising. It is shown that an integral character image is only possible to depict taking into consideration the specific features of characters’ idiolects, which help the reader to discern a character’s social status, age, educational background, gender, and emotional state. The aim of the article is to identify the challenges in rendering idiolects of literary characters, such as phonetic distortion of words, non-equivalent lexis, and non-standard syntax and to examine the relevant translation strategies and tactics of idiolect reproduction. Based on a comparative analysis of the original and translated texts, the research yielded a number of translation strategies such as the strategy of maximum preservation of idiolect characteristics and the strategy of partial preservation of idiolect characteristics. In the framework of the strategy of maximum preservation of idiolect characteristics, the tactic of parallel translation, the tactic of applying functional equivalents and the tactic of phonetic matching are singled out. In the framework of the strategy of partial preservation of idiolect characteristics, the tactics of compensation, omission, preservation, and substitution are singled out.

Factors modulating cross-linguistic co-activation in bilinguals

Activation of both of a bilingual’s languages during auditory word recognition has been widely documented. Here, we argue that if parallel activation in bilinguals is the result of a bottom-up process where phonetic features that overlap across two languages activate both linguistic systems, then the robustness of such parallel activation is in fact surprising. This is because phonemes across two different languages are rarely perfectly matched to each other in phonetic features. For instance, across Spanish and English, a “voiced” stop is realized in phonetically-distinct ways, and therefore, words that begin with voiced stops in English do not in fact fully overlap in phonetic features with words in Spanish. In two eye-tracking experiments using a visual world paradigm, we examined the effect of a phonemic match (English /b/ matched to Spanish /b/) vs. a phonetic match (English /b/ matched to Spanish /p/) on cross-linguistic co-activation (English words co-activating Spanish) in Spanish L1 and in Spanish L2 speakers. We found that while phonemic matching induced co-activation in both Spanish L1 and Spanish L2 speakers, phonetic matching did not. Together, these results indicate that co-activation of two languages in bilinguals may proceed through activation of categorical phonemic information rather than through activation of phonetic features.

SoundexHindi: A Phonetic Matching Algorithm for Hindi Written in English

Phonetic matching plays key role in retrieving information in environment using multiple Indic languages. The purpose of retrieval of information involves searching of information from large database with different perspective so that further analysis can be done. In order to achieve these techniques such as indexing can be used along with ranking of matching. The strings actually represent the same keyword however due to language complexity it is possible that the same word may be spelt differently by different persons. Since in rural or urban areas, the selected word may be spelt or pronounced either wrongly or can be spelt differently. In cases of exact matching of words may not return correct result rather different techniques associated with phonetic matching can be used for retrieval. In this paper we proposed an approach, which provides a simple and efficient way of matching the strings in Hindi written in Roman. Our approach works on text-conversion technique for any Indian languages, especially Hindi, Marathi etc.

Multilingual Entity Matching

The aim of this paper is to explore methods of multilingual entity matching. Name matching is currently the main technique used for entity resolution. When dealing with entities having features recorded in different languages and with different alphabets the basic approaches have serious limitation. The basic name matching approach using string comparison metrics is enriched with phonetic rules and with relational information. The results show that the approach using transliteration enhanced by phonetic matching provides with the best performance.

MetaSoundex Phonetic Matching for English and Spanish

Researchers confront major problems while searching for various kinds of data in large imprecise databases, as they are not spelled correctly or in the way they were expected to be spelled. As a result, they cannot find the word they sought. Over the years of struggle, pronunciation of words was considered as one of the practices to solve the problem effectively. The technique used to acquire words based on sounds is known as “Phonetic Matching”. Soundex was the first algorithm developed and other algorithms such as Metaphone, Caverphone, DMetaphone, Phonex etc., are also used for information retrieval in different environments. The main contribution of this paper is to analyze and implement the newly proposed MetaSoundex algorithm for fixing ill-defined data in English and Spanish languages. The newly developed MetaSoundex algorithm addresses the limitations of well-known phonetic matching techniques, Metaphone and Soundex. Specifically, the new algorithm provided results that are more accurate compared to both Soundex and Metaphone algorithms and has higher precision compared to Soundex, thus reducing the noise in the considered arena.

Reducing and Sustaining Duplicate Medical Record Creation by Usability Testing and System Redesign.

Duplicate medical record creation is a common and consequential health care systems error often caused by poor search system usability and inappropriate user training. We conducted two phases of scenario-based usability testing with patient registrars working in areas at risk of generating duplicate medical records. Phase 1 evaluated the existing search system, which led to system redesigns. Phase 2 tested the redesigned system to mitigate potential errors before health system-wide implementation. To evaluate system effectiveness, we compared the monthly potential duplicate medical record rates for preimplementation and postimplementation months. The existing system could not effectively handle a misspelling, which led to failed search and duplicate medical record creation. Using the existing system, 96% of registrars found commonly spelled patient names whereas only 69% successfully found complicated names. Registrars lacked knowledge and usage of a phonetic matching function to assist in misspelling. The new system consistently captured the correct patient regardless of misspelling, but search returned more potential matches, resulting in, on average, 4 seconds longer to select common names. Potential monthly duplicate medical record rate reduced by 38%, from 4% to 2.3% after implementation of the new system, and has sustained at an average of 2.5% for 2 years. Usability testing was an effective method to reveal problems and aid system redesign to deliver a more user friendly system, hence reducing the potential for medical record duplication. Greater standards for usability would ensure that these improvements can be realized before rather than after exposing patients to risks.

Need for Customized Soundex based Algorithm on Indian Names for Phonetic Matching

In any digitization program, the reproduction of the handwritten demographic data is a challenging job particularly for the records of previous decades. Nowadays, the requirement of the digitization of the individual’s past records becomes very much essential. In the areas like financial inclusion, border security, driving license, passport issuance, weapon license, banking sectors, health care and social welfare benefits, the individual’s earlier case history is a mandatory part of the decision making process. Documents are scanned and stored in a systematic method; each and every scanned document is tagged with a proper key. Documents are retrieved with the help of assigned key, for the purpose of data entry through the software program/ package. Here comes the difficulty that the data, particularly the critical personal data like name and father name etc., may not be legible for the reading purpose and the data entry operators type the characters as per their understanding. The chances of error is of high order in name variations in terms of duplicate characters, abbreviations, omissions, ignoring space between names and wrong spelling. Now the challenge is that, result of data retrieval over these key fields may not be proper because of the wrong data entry. We need to explore the opportunities and challenges for defining the effective strategies to execute this job without compromising the quality and quantity of the matches. In this scenario, we need to have an appropriate string matching algorithm with the phonetic matching. The algorithm is to be defined according to the nature, type and region of the data domain so that the search shall be phonetic based rather than simple string comparison. In this paper, I have tried to explain the need for customized soundex based algorithm on phonetic matching over the misspelt, incomplete, repetitive and partial prevalent data.

Phonetic Matching, Semanticized Phonetic Matching and Phono-Semantic Matching as Techniques in Keyword Selection

In foreign language vocabulary learning, the keyword method is among the most widely researched mnemonics and has been proved effective by numerous empirical studies. To use the keyword method, a keyword in the native language must be selected for the creation of an image or a sentence of the keyword “interacting” with the target word in the foreign language, thus facilitating retention and retrieval of the target word. In an attempt to contribute to successful application of the keyword method, this paper outlines three techniques in keyword selection, i.e., phonetic matching, semanticized phonetic matching, and phono-semantic matching, by mainly drawing on example pairs from Chinese and English. The phonetic and semantic links between keyword and target word pairs derived through each technique are analysed, and possible pitfalls are addressed as well.

Automating Corpora Generation with Semantic Cleaning and Tagging of Tweets for Multi-dimensional Social Media Analytics

Developing corpora from social media content involves convoluted cleaning. In this paper we propose and implement the automation of corpora building for facilitating social media mining and analytics. This automation process incorporates: a) metadata extraction and structuring b) semantic cleaning with tagging and c) learning domain terms/entities. The implementation performs comprehensive cleaning including abbreviation and slang correction, phonetic matching using metaphone algorithm, splitting joined words and identifying/learning entities. It identifies the entities, tags them and creates/updates a knowledgebase (KB) comprising of domain terms. The corpus thus constructed, facilitates multidimensional analysis and summarization. This proposed technique was tested with an experiment in which real world streaming tweets pertaining to Indian politics were collected, structured, cleaned and tagged. The results of the automation experiment can be stated as follows: a) the tweets although primarily in English, contained at times words from the regional languages. The algorithm does not recognize these words and they are construed as domain terms. An accuracy of 85.55% was achieved in identifying the correct domain terms and entities. b) The automation required human feedback and intervention which progressively reduced and reached a figure of 18% with the update and enhancement of the KB. This paper assumes relevance because the implementation automates the entire process of collecting and cleaning the tweets and yields a corpus suitable for multifaceted analysis.

Phonetic matching of auditory and visual speech develops during childhood: Evidence from sine-wave speech

The correspondence between auditory speech and lip-read information can be detected based on a combination of temporal and phonetic cross-modal cues. Here, we determined the point in developmental time at which children start to effectively use phonetic information to match a speech sound with one of two articulating faces. We presented 4- to 11-year-olds (N=77) with three-syllabic sine-wave speech replicas of two pseudo-words that were perceived as non-speech and asked them to match the sounds with the corresponding lip-read video. At first, children had no phonetic knowledge about the sounds, and matching was thus based on the temporal cues that are fully retained in sine-wave speech. Next, we trained all children to perceive the phonetic identity of the sine-wave speech and repeated the audiovisual (AV) matching task. Only at around 6.5years of age did the benefit of having phonetic knowledge about the stimuli become apparent, thereby indicating that AV matching based on phonetic cues presumably develops more slowly than AV matching based on temporal cues.

Improvement of Soundex Algorithm for Indian Language Based on Phonetic Matching

In a system with a large database, there always has been a problem that names may not be spe lled well or might not be spelled in a way thatone expected. So, data in the database gets degrad ed. In this case it is required to search the duplicates and merge them in the single entity. In doing so, one problem is that the way in which the strings would be compared. In such cases rather than looking for exact match, approximate string matching would be appreciable . One of the string matching techniques isPhonetic matching which is used to compare the name based on the pronunciation of the words. The similar sounding words could be retrieved from the large database using different phonetic matching algorithm and best known algorithm is Soundex algorithm. Phonetic matching is needed when many people from different culture come together. They either spe ak with different pronunciation or their writing habits are different. This scenario is very common in India, as we have many different languages like Hindi, Gujarati, Marathi, Tamil etc. In this research work Soundex algorithm is used for Hindi and Gujara ti language and applied on the names along with their variations in order to retrieve the output with minimum false hits.

Analysis and Comparative Study on Phonetic Matching Techniques

for names in large databases containing spelling variations has always been a problem. The first solution to the problem was proposed by Robert Russell in 1912 as he proposed SoundEx algorithm. SoundEx algorithm matches the names based on the sound of the words. The technique which is based on the pronunciation of the word is known as phonetic matching. SoundEx algorithm is one of the phonetic matching algorithms. We have discussed the other phonetic matching algorithm like edit distance algorithm, K-String and Q gram algorithm, Guth algorithm, Daitch Mokotoff algorithm and Metaphone coding algorithm. Our main focus is on the SoundEx algorithm and this paper also describes the issues with SoundEx algorithm.

Rule-Based Phonetic Matching Approach for Hindi and Marathi

Phonetic matching plays an important role in multilingual information retrieval, where data is manipulated in multiple languages and user need information in their native language which may be different from the language where data has been maintained. In such an environment, we need a system which matches the strings phonetically in any case. Once strings match, we can retrieve the information irrespective of languages. In this paper, we proposed an approach which matches the strings either in Hindi or Marathi or in cross-language in order to retrieve information. We compared our proposed method with soundex and Q-gram methods. We found better results as compared to these approaches.

DLworm: a system for workflow and repository management for digital libraries

This paper describes the Digital Library Workflow and Repository Management system (DLworm) to manage the repositories and to automate the workflow for digital library at any institution/organization. After a detailed and comprehensive analysis of the steps involved in maintaining a digital library system, a concrete model has been developed and tested. DLworm supports most of the functionalities required for such a large repository. DLworm is intended to formulate a fault tolerant system providing integrated environment for all users involved in the processes of maintaining and accessing a digital library. DLworm is a flexible and extensible system supporting phonetic matching for search queries. DLworm uses a three level hierarchy for maintaining the data in the repositories. We describe briefly the architecture that is robust and fulfils the demand for long term preservation of digital data. This article illustrates mechanisms for process management that makes DLworm highly efficient and user friendly.

Sounds like ‘Sutu’

Phonetic string matching involves taking two or more strings to determine if they match. Strings that are often orthographically spelt differently are matchable (nearly equivalent), either as people's names, e.g. Smith and Smythe, or as words, e.g. colour and color. Equivalent examples in various indigenous Southern African languages would be surnames, e.g. Radebe and Hadebe or words, e.g. mpša, mbwa, ntsha which are all equivalent words for ‘dog’; similarly, hlapi, tlhapi, inhlami for ‘fish’. The authors look at two common string matching algorithms, SoundEx and NYSIIS, and show how they are inadequate for Southern African use. Modifications of these algorithms for Southern African use as well as new algorithms are proposed.

A comparison of standard spell checking algorithms and a novel binary neural approach

In this paper, we propose a simple, flexible, and efficient hybrid spell checking methodology based upon phonetic matching, supervised learning, and associative matching in the AURA neural system. We integrate Hamming Distance and n-gram algorithms that have high recall for typing errors and a phonetic spell-checking algorithm in a single novel architecture. Our approach is suitable for any spell checking application though aimed toward isolated word error correction, particularly spell checking user queries in a search engine. We use a novel scoring scheme to integrate the retrieved words from each spelling approach and calculate an overall score for each matched word. From the overall scores, we can rank the possible matches. We evaluate our approach against several benchmark spellchecking algorithms for recall accuracy. Our proposed hybrid methodology has the highest recall rate of the techniques evaluated. The method has a high recall rate and low-computational cost.

Two‐month‐old infants match phonetic information in lips and voice

AbstractInfants aged 4.5 months are able to match phonetic information in the face and voice (Kuhl & Meltzoff, 1982; Patterson & Werker, 1999); however, the ontogeny of this remarkable ability is not understood. In the present study, we address this question by testing substantially younger infants at 2 months of age. Like the 4.5‐month‐olds in past studies, the 2‐month‐old infants tested in the current study showed evidence of matching vowel information in face and voice. The effect was observed in overall looking time, number of infants who looked longer at the match, and longest look to the match versus mismatch. Furthermore, there were no differences based on male or female stimuli and no preferences for the match when it was on the right or left side. These results show that there is robust evidence for phonetic matching at a much younger age than previously known and support arguments for either some kind of privileged processing or particularly rapid learning of phonetic information.

한영 혼용문에서 괄호 안 대역어구의 자동 인식

In Korean-English mixed documents, translated technical words are usually used with the attached full words or original words enclosed with parenthesis. In this paper, a collective method is presented to recognize and extract the translation phrases with using a base translation dictionary. In order to process the unregistered title words and translation words in the dictionary, a phonetic similarity matching method, a translation partial matching method, and a compound word matching method are newly proposed. The experiment result of each method was measured in F-measure(the alpha is set to 0.4) ; exact matching of dictionary terms as a baseline method showed 23.8%, the hybrid method of translation partial matching and phonetic similarity matching 75.9%, and the compound word matching method including the hybrid method 77.3%, which is 3.25 times better than the baseline method.