Fuzzy String Matching Research Articles

Using of Fuzzy Logic Matching Algorithm for Hindi Dataset

Various new computing methods based on Fuzzy Logic can be used in various systems. Numerous researchers and developers are working with Fuzzy Logic and its Applications. One such application is applied on the dataset which contains the values in Hindi Language; the major problem is that data come from different sources and when do the merge operation it will create a lot of problem; since same names are represented by different ways of writing it. Here, we apply a technique of fuzzy matching algorithm for hindi dataset. We are applying for the project of sharing bike concept which comes from different data sources. Fuzzy Logic is a very promising mathematical approach for training the dataset which being characterized by subjectivity & imprecision. More emphasis is put on the fuzzy string matching and its criteria for the performing the merge operation. This paper presents an analysis of the results achieved using fuzzy logic to model the dataset. Assessing the performance of datasets after performing matching algorithm at different levels using traditional methods.

Toward Effective Aircraft Call Sign Detection Using Fuzzy String-Matching between ASR and ADS-B Data

Recently, artificial intelligence and data science have witnessed dramatic progress and rapid growth, especially Automatic Speech Recognition (ASR) technology based on Hidden Markov Models (HMMs) and Deep Neural Networks (DNNs). Consequently, new end-to-end Recurrent Neural Network (RNN) toolkits were developed with higher speed and accuracy that can often achieve a Word Error Rate (WER) below 10%. These toolkits can nowadays be deployed, for instance, within aircraft cockpits and Air Traffic Control (ATC) systems in order to identify aircraft and display recognized voice messages related to flight data, especially for airports not equipped with radar. Hence, the performance of air traffic controllers and pilots can ultimately be improved by reducing workload and stress and enforcing safety standards. Our experiment conducted at Tangier’s International Airport ATC aimed to build an ASR model that is able to recognize aircraft call signs in a fast and accurate way. The acoustic and linguistic models were trained on the Ibn Battouta Speech Corpus (IBSC), resulting in an unprecedented speech dataset with approved transcription that includes real weather aerodrome observation data and flight information with a call sign captured by an ADS-B receiver. All of these data were synchronized with voice recordings in a structured format. We calculated the WER to evaluate the model’s accuracy and compared different methods of dataset training for model building and adaptation. Despite the high interference in the VHF radio communication channel and fast-speaking conditions that increased the WER level to 20%, our standalone and low-cost ASR system with a trained RNN model, supported by the Deep Speech toolkit, was able to achieve call sign detection rate scores up to 96% in air traffic controller messages and 90% in pilot messages while displaying related flight information from ADS-B data using the Fuzzy string-matching algorithm.

Rancang Bangun Chatbot Pengaduan Kekerasan Perempuan Anak Dengan Metode Fuzzy String Matching Dan Enhanced Confix Stripping Stemmer

Violence against women is a serious issue affecting many individuals in society. In Indonesia, including Kota Probolinggo, the number of cases of violence against women remains high. This research aims to design a chatbot application as a means for reporting violence against women and children (VAWC) in Kota Probolinggo. The application provides an anonymous and safe space for victims to report their complaints without revealing their identities. Additionally, the chatbot delivers information and education about different types of violence against women and preventive measures. The development of the chatbot application utilizes Fuzzy String Matching, Enhanced Confix Stripping Stemmer, and preprocessing methods to enhance data search accuracy and provide appropriate responses to users. The testing results show that the chatbot offers responses with satisfactory accuracy. Leveraging AI technology in the form of a chatbot is expected to provide an effective solution in addressing and preventing violence against women, offering a secure and anonymous platform for victims to report incidents. The application is envisioned to offer better support and assistance to victims of violence.
 The chatbot application implements fuzzy string matching, the Levenshtein distance algorithm, and Enhanced Confix Stripping Stemmer to process user input before matching it with the database. During the Enhanced Confix Stripping Stemmer process, sentences are transformed into an array of basic words, enhancing the accuracy of data matching when using fuzzy string matching and the Levenshtein distance algorithm. Testing results indicate a 4.77% difference between manual calculations and the chatbot system. This difference may be attributed to variations in text processing and other factors.

Hybridizing Fuzzy String Matching and Machine Learning for Improved Ontology Alignment

Ontology alignment has become an important process for identifying similarities and differences between ontologies, to facilitate their integration and reuse. To this end, fuzzy string-matching algorithms have been developed for strings similarity detection and have been used in ontology alignment. However, a significant limitation of existing fuzzy string-matching algorithms is their reliance on lexical/syntactic contents of ontology only, which do not capture semantic features of ontologies. To address this limitation, this paper proposed a novel method that hybridizes fuzzy string-matching algorithms and the Deep Bidirectional Transformer (BERT) deep learning model with three machine learning regression classifiers, namely, K-Nearest Neighbor Regression (kNN), Decision Tree Regression (DTR), and Support Vector Regression (SVR), to perform the alignment of ontologies. The use of the kNN, SVR, and DTR classifiers in the proposed method resulted in the building of three similarity models (SM), encoded SM-kNN, SM-SVR, and SM-DTR, respectively. The experiments were conducted on a dataset obtained from the anatomy track in the Ontology Alignment and Evaluation Initiative 2022 (OAEI 2022). The performances of the SM-kNN, SM-SVR, and SM-DTR models were evaluated using various metrics including precision, recall, F1-score, and accuracy at thresholds 0.70, 0.80, and 0.90, as well as error rates and running times. The experimental results revealed that the SM-SVR model achieved the best recall of 1.0, while the SM-DTR model exhibited the best precision, accuracy, and F1-score of 0.98, 0.97, and 0.98, respectively. Furthermore, the results showed that the SM-kNN, SM-SVR, and SM-DTR models outperformed state-of-the-art alignment systems that participated in the OAEI 2022 challenge, indicating the superior capability of the proposed method.

Pipeline for the Automatic Extraction of Procedural Knowledge from Assembly Instructions into Controlled Natural Language

This paper presents the application of a Natural Language Processing (NLP) pipeline, which automatically extracts procedural knowledge in a standardized way from assembly instructions. The developed pipeline is able to parse and process written German assembly instructions regardless of the language discourse. The pipeline helps resolve ambiguities in assembly instructions by converting them into a Controlled Natural Language (CNL). The pipeline fully automates the translation process from free-text assembly instructions to CNL representations. We investigated and evaluated the efficiency and robustness of the NLP pipeline along multiple dimensions, such as different assembly process designers, language and fuzzy string matching models. To test the developed pipeline we used to automatically extract procedural knowledge in a standardized way for 2,740 assembly instructions obtained from automotive industry. Our investigation shows that the NLP pipeline is able to extract CNL representations with high accuracy (ø 87%). Downstream applications, such as assembly line balancing, can reuse the uniformly extracted procedural knowledge.

More extreme duplication in FDA Adverse Event Reporting System detected by literature reference normalization and fuzzy string matching.

Literature reports of adverse drug events can be replicated across multiple companies, resulting in extreme duplication (defined as a majority of reports being duplicates) in the FDA Adverse Event Reporting System (FAERS) database because they can escape legacy duplicate detection algorithms routinely deployed on that data source. Literature reference field, added to in 2014, could potentially be utilized to identify replicated reports. FAERS does not enforce adherence to the Vancouver referencing convention, thus the same article may be referenced differently leading to duplication. The objective of this analysis is to determine if variations of the same literature references observed in FAERS can be resolved with text normalization and fuzzy string matching. We normalized the literature references recorded in the FAERS database through the first quarter of 2021 with a rule-based algorithm so that they better conform to the Vancouver convention. Levenshtein distance was then utilized to merge sufficiently similar normalized literature references together. Normalization of literature references increases the percentage that can be parsed into author, title, and journal from 61.74% to 93.93%. We observe that about 98% of pairs within groups do have a Levenshtein similarity of the title above the threshold. The extreme duplication ranged from 66% to 87% with a median of 72% of reports being duplicates and often involved addictovigilance scenarios. We have shown that these normalized references can be merged via fuzzy string matching to improve enumeration of all the individual case safety reports that refer to the same article. Inclusion of the PubMed ID and adherence to the Vancouver convention could facilitate identification of duplicates in the FAERS dataset. Awareness of this phenomenon may improve disproportionality analysis, especially in areas such as addictovigilance.

The burden of geriatric traumatic brain injury on trauma systems: Analysis of 348,800 Medicare inpatient claims.

Traumatic brain injury (TBI) is a leading cause of death and disability in older adults. The aim of this study was to characterize the burden of TBI in older adults by describing demographics, care location, diagnoses, outcomes, and payments in this high-risk group. Using 2016-2019 Centers for Medicare & Medicaid Services (CMS) Inpatient Standard Analytical Files (IPSAF), patients >65 years with TBI (>1 injury ICD-10 starting with "S06") were selected. Trauma center levels were linked to the IPSAF file via American Hospital Association Hospital Provider ID and fuzzy-string matching. Patient variables were compared across trauma center levels. Three hundred forty-eight thousand eight hundred inpatients (50.4% female; 87.1% white) from 2963 US hospitals were included. Level I/II trauma centers treated 66.9% of patients; non-trauma centers treated 21.5%. Overall inter-facility transfer rate was 19.2%; in Level I/II trauma centers transfers-in represented 23.3% of admissions. Significant TBI (Head AIS ≥3) was present in 70.0%. Most frequent diagnoses were subdural hemorrhage (56.6%) and subarachnoid hemorrhage (30.6%). Neurosurgical operations were performed in 10.9% of patients and operative rates were similar regardless of center level. Total unadjusted mortality for the sample was 13.9%, with a mortality of 8.1% for those who expired in-hospital, and an additional 5.8% for those discharged to hospice. Medicare payments totaled $4.91B, with the majority (73.4%) going to Level I/II trauma centers. This study fills a gap in TBI research by demonstrating that although the majority of older adult TBI patients in the United States receive care at Level I/II trauma centers, a substantial percentage are managed at other facilities, despite 1 in 10 requiring neurosurgical operation regardless of level of trauma center. This analysis provides preliminary data on the function of regionalized trauma care for older adult TBI care. Future studies assessing the efficacy of early care guidelines in this population are warranted.

Concurrent 22. Presentation for: Extracting events from Daily Drilling Reports using Fuzzy String Matching

Presented on Thursday 19 May: Session 22 Continuous monitoring of oil drilling operations reduces process interruptions and equipment failure. It also contributes to the development of Key Performance Indicators, which leads to more efficient resource management. Daily Drilling Reports (DDRs) have long been the primary way of recording noticeable events, such as stuck pipe. DDRs came to constitute a valuable information base for most oil drilling companies. However, the task of extracting knowledge from DDRs can be costly and time-consuming. This work proposes an approach to recognise drilling events in DDRs using a rule-based language processing method called Fuzzy String Matching (FSM). We applied the FSM algorithm to search for a set of predefined keywords and key phrases to extract possible Invisible Lost Time (ILT) events from DDRs that may indicate risks or low operational efficiency. The fuzzy part of the algorithm allows the identification of terms or expressions that match the pre-established ones approximately rather than exactly, accounting for typos and different suffixes or prefixes. The proposed solution was applied on a data set of 392 real-world DDR records from a drilling company using a set of six ILT event’s key phrases annotated by Subject Matter Specialists. This process can be readily replicated to other events. The results show that in 116 reports tagged as normal, 92 records were identified as possible ILT events, which represents, in hours, 56% of the total drill normal time. Such promising results can lead to very significant improvements in identifying and extracting drilling events within DDRs. To access the presentation click the link on the right. To read the full paper click here

Extracting events from Daily Drilling Reports using Fuzzy String Matching

Continuous monitoring of oil drilling operations reduces process interruptions and equipment failure. It also contributes to the development of Key Performance Indicators, which leads to more efficient resource management. Daily Drilling Reports (DDRs) have long been the primary way of recording noticeable events, such as stuck pipe. DDRs came to constitute a valuable information base for most oil drilling companies. However, the task of extracting knowledge from DDRs can be costly and time-consuming. This work proposes an approach to recognise drilling events in DDRs using a rule-based language processing method called Fuzzy String Matching (FSM). We applied the FSM algorithm to search for a set of predefined keywords and key phrases to extract possible Invisible Lost Time (ILT) events from DDRs that may indicate risks or low operational efficiency. The fuzzy part of the algorithm allows the identification of terms or expressions that match the pre-established ones approximately rather than exactly, accounting for typos and different suffixes or prefixes. The proposed solution was applied on a data set of 392 real-world DDR records from a drilling company using a set of six ILT event’s key phrases annotated by Subject Matter Specialists. This process can be readily replicated to other events. The results show that in 116 reports tagged as normal, 92 records were identified as possible ILT events, which represents, in hours, 56% of the total drill normal time. Such promising results can lead to very significant improvements in identifying and extracting drilling events within DDRs.

The public health burden of geriatric trauma: Analysis of 2,688,008 hospitalizations from Centers for Medicare and Medicaid Services inpatient claims.

Geriatric trauma care (GTC) represents an increasing proportion of injury care, but associated public health research on outcomes and expenditures is limited. The purpose of this study was to describe GTC characteristics, location, diagnoses, and expenditures. Patients at short-term nonfederal hospitals, 65 years or older, with ≥1 injury International Classification of Diseases, Tenth Revision, were selected from 2016 to 2019 Centers for Medicare and Medicaid Services Inpatient Standard Analytical Files. Trauma center levels were linked to Inpatient Standard Analytical Files data via American Hospital Association Hospital ID and fuzzy string matching. Demographics, care location, diagnoses, and expenditures were compared across groups. A total of 2,688,008 hospitalizations (62% female; 90% White; 71% falls; mean Injury Severity Score, 6.5) from 3,286 hospitals were included, comprising 8.5% of all Medicare inpatient hospitalizations. Level I centers encompassed 7.2% of the institutions (n = 236) but 21.2% of hospitalizations, while nontrauma centers represented 58.5% of institutions (n = 1,923) and 37.7% of hospitalizations. Compared with nontrauma centers, patients at Level I centers had higher Elixhauser scores (9.0 vs. 8.8) and Injury Severity Score (7.4 vs. 6.0; p < 0.0001). The most frequent primary diagnosis at all centers was hip/femur fracture (28.3%), followed by traumatic brain injury (10.1%). Expenditures totaled $32.9 billion for trauma-related hospitalizations, or 9.1% of total Medicare hospitalization expenditures and approximately 1.1% of the annual Medicare budget. The overall mortality rate was 3.5%. Geriatric trauma care accounts for 8.5% of all inpatient GTC and a similar percentage of expenditures, the most common injury being hip/femur fractures. The largest proportion of GTC occurs at nontrauma centers, emphasizing their vital role in trauma care. Public health prevention programs and GTC guidelines should be implemented by all hospitals, not just trauma centers. Further research is required to determine the optimal role of trauma systems in GTC, establish data-driven triage guidelines, and define the impact of trauma centers and nontrauma centers on GTC mortality. Therapeutic/care management, Level III.

Efficient processing of continuous spatial-textual queries over geo-textual data stream

&lt;p&gt;Due to the extensive use of social media and mobile devices, unbounded and massive data is generated continuously. The need to process this big data is increasing day by day. The traditional data processing algorithms fail to cater to the need of processing data generated by various applications such as digital geo-based advertising, and recommendation systems. There has been a high demand to process continuous spatial fuzzy textual queries over data stream of spatial-textual objects with high density by present locationbased and social network-based service applications. For the spatialkeyword data stream, the performance plays a vital role as the geo information and keyword description matching is needed for every incoming streaming object. The various continuous geo-keyword query processing methods normally lack the support for fuzzy keyword matching when processing the objects from the geo-textual data stream. The edit distancebased approach with the adaptive partitioning tree index for the queries is used for fuzzy string matching and it outperforms than the existing approaches in storage cost and query performance cost.&lt;/p&gt;

&amp;lt;i&amp;gt;dh2loop&amp;lt;/i&amp;gt; 1.0: an open-source Python library for automated processing and classification of geological logs

Abstract. A huge amount of legacy drilling data is available in geological survey but cannot be used directly as they are compiled and recorded in an unstructured textual form and using different formats depending on the database structure, company, logging geologist, investigation method, investigated materials and/or drilling campaign. They are subjective and plagued by uncertainty as they are likely to have been conducted by tens to hundreds of geologists, all of whom would have their own personal biases. dh2loop (https://github.com/Loop3D/dh2loop, last access: 30 September 2021​​​​​​​) is an open-source Python library for extracting and standardizing geologic drill hole data and exporting them into readily importable interval tables (collar, survey, lithology). In this contribution, we extract, process and classify lithological logs from the Geological Survey of Western Australia (GSWA) Mineral Exploration Reports (WAMEX) database in the Yalgoo–Singleton greenstone belt (YSGB) region. The contribution also addresses the subjective nature and variability of the nomenclature of lithological descriptions within and across different drilling campaigns by using thesauri and fuzzy string matching. For this study case, 86 % of the extracted lithology data is successfully matched to lithologies in the thesauri. Since this process can be tedious, we attempted to test the string matching with the comments, which resulted in a matching rate of 16 % (7870 successfully matched records out of 47 823 records). The standardized lithological data are then classified into multi-level groupings that can be used to systematically upscale and downscale drill hole data inputs for multiscale 3D geological modelling. dh2loop formats legacy data bridging the gap between utilization and maximization of legacy drill hole data and drill hole analysis functionalities available in existing Python libraries (lasio, welly, striplog).

Food Ingredients Similarity Based on Conceptual and Textual Similarity

Open Food Facts provides a database of food products such as product names, compositions, and additives, where everyone can contribute to add the data or reuse the existing data. The open food facts data are dirty and needs to be processed before storing the data to our system. To reduce redundancy in food ingredients data, we measure the similarity of ingredient food using two similarities: the conceptual similarity and textual similarity. The conceptual similarity measures the similarity between the two datasets by its word meaning (synonym), while the textual similarity is based on fuzzy string matching, namely Levenshtein distance, Jaro-Winkler distance, and Jaccard distance. Based on our evaluation, the combination of similarity measurements using textual and Wordnet similarity (conceptual) was the most optimal similarity method in food ingredients.

Adaptive Fuzzy String Matching: How to Merge Datasets with Only One (Messy) Identifying Field

AbstractA single dataset is rarely sufficient to address a question of substantive interest. Instead, most applied data analysis combines data from multiple sources. Very rarely do two datasets contain the same identifiers with which to merge datasets; fields like name, address, and phone number may be entered incorrectly, missing, or in dissimilar formats. Combining multiple datasets absent a unique identifier that unambiguously connects entries is called the record linkage problem. While recent work has made great progress in the case where there are many possible fields on which to match, the much more uncertain case of only one identifying field remains unsolved: this fuzzy string matching problem, both its own problem and a component of standard record linkage problems, is our focus. We design and validate an algorithmic solution called Adaptive Fuzzy String Matching rooted in adaptive learning, and show that our tool identifies more matches, with higher precision, than existing solutions. Finally, we illustrate its validity and practical value through applications to matching organizations, places, and individuals.

Exploring the variable efficacy of Google speech-to-text with spontaneous bilingual speech in Cantonese and English

With the growth of Automatic Speech Recognition (ASR) and voice user interface software, it is important to test for efficacy across different language varieties and identify sources of bias. Recent work assessing ASR efficacy and bias implicates factors like race, gender, dialect, and age as leading to different efficacy rates. Multilingualism presents another source of variation that ASR systems must grapple with, ranging from code-switching to phonetic variation both within and across speakers. Thus, variable ASR performance is likely exacerbated for multilingual communities. Using a spontaneous Cantonese-English bilingual speech corpus (Johnson, 2021), this study tests the efficacy of Google Speech-To-Text (STT) language models (Canadian English and Hong Kong Cantonese) with a heterogeneous bilingual speech community. Efficacy is assessed via fuzzy string matching between the STT and the manually corrected transcripts. STT performance is variable but overall better for English. Confidence ratings and matching scores are evaluated alongside listener ratings of perceived accentedness and various demographic groupings. The results of this study will provide practical guidance for using STT in the context of speech production research pipelines while highlighting its drawbacks concerning bias in a relatively understudied, multilingual group.

Fun and Easy Georeferencing with a New Online Tool from South Africa

A new online tool for georeferencing specimen records has been developed through the Natural Science Collections Facility (NSCF) in South Africa to address the need for rapid, high quality georeferencing of specimen collections in the region (Fig. 1). A dataset of specimen records with Darwin Core fields dwc:scientificName, dwc:country, optional dwc:stateProvince, dwc:locality or dwc:verbatimLocality, optional dwc:recordedBy, and a record identifier such as dwc:occurrenceID (see dwc.tdwg.org/terms for definitions) is first uploaded into the tool and a team of georeferencers then work to georeference the dataset. Fuzzy string matching is used to group similar locality strings and to search for potential matching georeferences from a georeference database. The tool aims to improve efficiency by storing georeferenced localities so that they can be reused when the same locality is encountered again in other datasets. Thus, a locality only needs to be georeferenced once, and that georeference is reused for any other permutations of that locality string. A georeference includes the most important metadata from the Darwin Core standard: a measure of uncertainty, dwc:georeferenceDate, dwc:georeferencedBy, dwc:georeferenceProtocol, dwc:georeferenceSources, and the all too often neglected dwc:geodeticDatum. dwc:georeferencedByID is included for recording the ORCID iD of the georeferencer to facilitate attribution further down the data publication pipeline. In theory, the process of georeferencing should become more efficient with time as the georeference database grows. The georeferencing process is gamified by showing each georeferencer their own numbers of georeferenced records as they work, and they can see activity of fellow georeferencers as the dataset statistics update in real time. Dataset owners can also see overall progress with the dataset and numbers of records georeferenced by each team member, which may be useful for management purposes. Once a dataset is completed, it is downloaded with the new georeferences so these can be incorporated back into the original source database. Within the landscape of currently available georeferencing tools the system presented here is specifically placed to facilitate the management of the georeferencing process for a dataset by a team of georeferencers. The georeferencing workflow still requires a full suite of tools for finding coordinates for localities, such as a GIS, gazetteers and online resources, as well as a specific georeferencing protocol. It essentially replaces the use of spreadsheets for doing georeferencing, or doing georeferences directly in a collection database, which can be inefficient. Related to this, it includes a quality assurance process whereby georeferences are checked for correctness and adherence to the protocol being used, and for identifying geographic and environmental outliers for each species within the dataset. In this way the tool supports current workflows and best practices for georeferencing (e.g. Chapman and Wieczorek (2020), Zermoglio et al. (2020)). The technology stack includes Firebase as the primary database, ElasticSearch for fuzzy string matching, and the user interface is built with the modern Javascript framework Svelte. The tool has been in use by the NSCF since April 2021 after being populated with approximately 300 000 existing georeferences for southern Africa from various sources, including the South African National Biodiversity Institute (SANBI) Gazetteer and several collections databases. While initial emphasis in developing the tool has focussed on southern Africa, the tool can be extended to other regions easily. Please contact data@nscf.org.za for further information.

DETEKSI KEMIRIPAN ARTIKEL MELALUI KEYWORDS DENGAN METODE FUZZY STRING MATCHING DALAM NATURAL LANGUAGE PROCESSING

Natural Language Processing (NLP) is part of artificial intelligence that focuses on natural language processing. That is the language commonly used by humans in communicating with each other. In writing, articles can also be applied. To find out the similarity of the contents of a scientific article to another will make it easier for readers to make selections and make it easier to collect similar documents. Likewise with differences in article content even though they use the same keywords. Measurements are made using the keywords contained in the abstract which consists of several words and the number of keywords. The method used is to use fuzzy string matching to get the number of keywords used in an article. Then the calculation will be done for each keyword used and will be sorted by priority according to the position of the keyword. The search will be carried out starting from the title, abstract, keywords, content, and references. The number of keywords found in each article is shared across all keywords found to generate a similarity percentage level. If the results found are the same in number from one article to another, it can even be categorized that the contents of the article have similar content. The testing process is carried out by counting the number of keywords in the source then comparing all the source keys to the destination articles in the database by searching and comparing each word.

Using fuzzy string matching for automated assessment of listener transcripts in speech intelligibility studies

Many studies of speech perception assess the intelligibility of spoken sentence stimuli by means of transcription tasks (‘type out what you hear’). The intelligibility of a given stimulus is then often expressed in terms of percentage of words correctly reported from the target sentence. Yet scoring the participants’ raw responses for words correctly identified from the target sentence is a time-consuming task, and hence resource-intensive. Moreover, there is no consensus among speech scientists about what specific protocol to use for the human scoring, limiting the reliability of human scores. The present paper evaluates various forms of fuzzy string matching between participants’ responses and target sentences, as automated metrics of listener transcript accuracy. We demonstrate that one particular metric, the token sort ratio, is a consistent, highly efficient, and accurate metric for automated assessment of listener transcripts, as evidenced by high correlations with human-generated scores (best correlation: r = 0.940) and a strong relationship to acoustic markers of speech intelligibility. Thus, fuzzy string matching provides a practical tool for assessment of listener transcript accuracy in large-scale speech intelligibility studies. See https://tokensortratio.netlify.app for an online implementation.

ClimActor, harmonized transnational data on climate network participation by city and regional governments

Cities and regions have become increasingly engaged in global climate change governance. They are pledging their own climate mitigation targets and participating in membership networks that typically are transnational in nature and engage thousands of subnational governments. Researching these growing trends in participation has been difficult due to the disparate and inconsistent nature of this self-reported data. To facilitate future analyses of these actors, we introduce ClimActor, the largest harmonized global dataset of more than 10,000 city and regional governments participating in networks like the Global Covenant of Mayors for Climate and Energy, C40 Cities for Climate Leadership, ICLEI Local Leaders for Sustainability, among others. We include key contextual information on each actor’s population, geographic location, and administrative jurisdiction to facilitate disambiguation of potential overlaps in actions or emissions. We also provide a series of cleaning functions based on phonetic and fuzzy string matching algorithms within an open-source R package to make it easy for anyone to immediately use the ClimActor dataset with other relevant data.

Analysis and safety engineering of fuzzy string matching algorithms

In this paper we explore fuzzy string matching in an automatic ticket classification and processing system. We compare performance of the following string similarity algorithms: Longest Common Subsequence (LCS), Dice coefficient, Cosine Similarity, Levenshtein (edit) distance and Damerau distance. Through optimisation, we accomplished a 15% improvement in the ratio of false positives to true positive classifications over the existing approach used by a customer support system for free customers. To introduce greater safety; we compliment fuzzy string matching algorithms with a second layer Convolutional Neural Network (CNN) binary classifier, achieving an improved keyword classification ratio for two ticket categories by a relative 69% and 78%. Such an approach allows for classification to only be applied where a desired level of safety achieved, such as in instances where automated answers.