Search

YouTube is one of the biggest platforms where people around the world can share informative videos. In YouTube, we can find the code-switching and code-mixing phenomenon easily. There are two dominant types of code-switching and code-mixing. The first type is code-switching and code-mixing used by Group 1: Indonesian YouTubers who live abroad, while the second type is code-switching and code-mixing used by Group 2: Indonesian YouTubers who live in the country. This study aimed to find out which group switched and mixed language the most. The transcripts of the chosen YouTube video were the collected data. The qualitative method was used in this study, and the data were analyzed using document analysis based on the framework of Myers-Scotton (1993) and Muysken (2000). The result showed that Group 1 used code-switching and code-mixing the most, with a total of 288 pieces of data. Group 2 used code-switching and code-mixing with a total of 262 pieces of data. The most dominant type of code-switching used by both groups was intra-sentential, a total of 154 pieces of data. The less dominant type of code-mixing used was insertion, a total of 253 pieces of data.

This descriptive qualitative research essentially aimed to find out the types of code mixing on Brown Sugar Battle video. Titan Tyra published the video on June 17th, 2019. The researchers chose the utterances of Titan Tyra that contained code mixing as the data source of this research. In collecting the data, the researchers conducted observational method and non-participatory technique. Additionally, in analyzing the data, the researchers applied bilingual-differentiating identification method and differentiating comparative link technique. Furthermore, the researcher applied the theory of code mixing proposed by Muysken (2000). The result revealed that the speaker frequently inserted English words and phrases in the utterance of Indonesian language. The speaker mixed language to do the review of brown sugar drink. From 53 data, there were 31 data of insertion and 22 data of congruent lexicalization. The type of insertion became the most frequently uttered type. Also, the speaker did not utter English clause in utterances.

The study was conducted in order to describe the forms and the functions of code-switching and code-mixing in drama performance of ALE of the students of ICP batch 2017. The study was descriptive qualitative research. The data are collected by observation and documentation in the drama performance in Art and Language Exhibition ‘Prahara Ing Argabelah’. The finding shows that 24 unit analysis of code-switching found 6 used tag-switching, 15 used inter-sentential switching, and 3 used intra-sentential switching. Furthermore, 27 unit analysis of code-mixing performed 21 intra-sentential code mixing, 5 intra-lexical code mixing, and 1 involving a change of pronunciation. Then, from 24 unit analysis of code-switching, the researcher found six functions in the drama conversation. There are 6 addressee specifications, 2 repetitions, 7 interjections, 1 message qualification, 2 personalization and objectivizations, and 6 facilities of expression. Moreover, the researcher just found 3 functions from 27 unit analysis of code-mixing. There are 1 repetition, 2 message qualifications, and 24 facilities of expression.

The phenomenon of mixing and switching languages can easily be found both in real life and on social media, one of which is on Instagram. On Instagram, someone can express themselves in the Instagram caption as Jun Bintang did. Jun Bintang is a multilingual who uses Balinese, Indonesian and English as code mixing and code switching on Instagram social media. Investigating code mixing and code switching as found in the captions of Jun Bintang posts on his Instagram account was the sole research objective. Qualitative research design with descriptive method was applied on code mixing and code switching phenomena. The research objects were focused on Jun Bintang’s posts in three languages, they were, Balinese, Indonesian, and English. Data were collected through observations conducted in March to May,2020. Results of data analysis showed that code mixing types frequently used were such types like 1) insertion, 2) congruent lexicalization, and 3) alternation. Whilst, code switching types used were that of 1) inter-sentential, and 2) intra-sentential. Reasons for using code mixing and switching were identified as showing 1) empathy, 2) ease of using first language, 3) purpose of talking specific topic, 4) prestige, 5) interjection, 6) motive, and 7) group identity.

This article attempts a comparative analysis of code-switching and code-mixing in the Nigerian music industry, using the lyrics of Flavour and 9ice as a case study. Although the English language is the national language in Nigeria and the language used by most of the musicians for the composition of their songs, and due to the linguistic plurality of Nigeria, most of these musicians tend to lace their songs chunks of words and phrases from their mother tongue or at least one of the three major languages in Nigeria, which are Hausa, Igbo, and Yoruba. The Markedness Model by Myers-Scotton (1993) is used as the framework to interrogate the switching and mixing in the codes used by these selected musicians and we find that while most code-switching is done in three languages – English, Nigerian Pidgin and the artist’ first language (mother tongue) – their mother tongue plays the prominent role. Code-switching or code-mixing in these songs, therefore, becomes a depiction of the Nigerian state with its diverse languages and it provides the links between the literates and the illiterates thereby giving the artiste the popularity desired. The study concludes that the unique identity created by code-switching and code-mixing in the Nigerian music industry has a positive influence on music lovers, helping artists to achieve wide patronage and reflecting the ethnolinguistic diversity of the Nigerian nation.

The objectives of this study are (1). To find the types of code mixing used in the novel Magic Hour. (2). To know the Indonesian meaning of code mixing used in the novel Magic Hour. (3). To know the most typical type code mixing used in the novel Magic Hour. This research uses descriptive qualitative method to collect the data. The data are the English words. The writer collects data by finding the types of Indonesian English code mixing, selecting sentence that consist of Indonesian-English code mixing. The analysis shows that there are some types that use, such as: Intra sentential mixing, intra lexical code mixing, involving a change of pronunciation. The types of code mixing which found in the novel are selected into the most typical forms. The writer looked for Indonesian meaning of code mixing. The achievements of this research are expected to be additional source of code mixing study and it is meant to conduct and to develop in the code mixing field.

This research concerned with the types of code switching and code mixing used by the teachers, teachers’ reasons in using code switching and code mixing, and students’ perception of code switching and code mixing used by the teacher in the class. This research was a descriptive qualitative research. The researcher used observation checklist, video recording, interview, and questionnaire as the instruments of this research. The data of this research was analyzed by classifying the types of code switching used theory of Poplack and the types code mixing used theory of Hoffman. The result of this research shows that the English teachers did all the types of code switching in English class. Namely; tag code switching, inter-sentential code switching, and intra-sentential code switching. Meanwhile, the researcher found that the teachers only used two from the three types of code mixing. Namely; intra-sentential code mixing and intra-lexical code mixing. The type that did not appear was involvement in a change of pronunciation. This study revealed that the teachers’ reason in using code switching and code mixing are; to help the students in learning English and to make the students more focus in learning process. For the students’ perception, most of the students agreed that the use of code switching and code mixing helps them understand the materials in English language learning class.

Code-mixing and code-switching are the marketing phenomenon researches have been working for around decades. While different types and their properties for code-mixing and code-switching have been well studied for most of the linguistic varieties as a marketing tool, very less is documented about how they are used as communicative and persuasive tool in advertisements. This analysis is an attempt to overcome with this gap by analyzing the level and describing the current phenomenon of English code-mixing and code-switching in Indian cold drink advertisements. It is noticed that the use of English codes in Hindi advertisements or even in few of complete English advertisements made specifically for Indian television is different from other genres of communication since they are creatively utilized to produce persuasive and convincing effects. The paper investigated the level of such mixing in Indian advertisements basically shown in Hindi speaking area, what kinds of English units are in use as well as certain other aspects related to English mixing which are employed as a persuasive device for the advertisements . English lexical mixing were found to be common English code-mixing elements, while simple sentences, out of all types of sentences, were the most frequent items appearing at the level of code-switching.

This study uses sociopragmatic theory, especially regarding code-mixing in speech acts at opening an event, greeting the listeners, and talking to interviewees. The purpose of writing is to determine the form of code-mixing and explain the meaning of the code-mixing used by the radio announcer, whether it is entered into locus, illocution, or perlocution. This paper is divided into three stages. The first stage is to provide data using the listening method by listening to radio broadcasts followed by the recording method. At the analysis stage, the data obtained were analyzed using related theories. At the last stage, it is making the conclusions and suggestions. In this research, the average Classy FM radio announcers use code-mixing in their speech acts when broadcasting. Apart from being influenced by the names of the programs that is titled in English, it is also influenced by the standardization of the words used by Classy FM itself. For example, to invite the listeners to keep listening, the term”keep in tune" is used. The term of Classy FM listeners refers to the term ”Classy People," and many other terms are used. Besides, it is also found that in one speech act, there are three types of actions that are included in it, namely the act of locus, illocution, and perlocution. Although there are also some data found only in the locus (speech acts to state something).

Sociolinguistics is a study of the connection between language and society. People have a different language style when they interact with each one. By many variations that they have, it can be possible for them to mix their language in their utterance—mixing one language with the other languages, in the sociolinguistics field, is called by code-mixing. It has been common to use of code-mixing in society. It happens in the one of the public figures and famous that is Mr. Nadiem Makariem. Therefore, this research focuses on code mixing that emerges on video of Mr. Nadiem Makariem. The objective of this research is to find out the types and levels of code mixing that appear on the video. This research was descriptive qualitative method and the researchers act as the main instrument of this research. In collecting the data, this research utilized documentation method. This study employed content analysis focusing on analyzing the types of code mixing which defined by Hoffman and the levels of code mixing that argued by Suwito. Then, the result of types and levels of code mixing were counted by using Walizer’s formula. The result shows that there were 134 data in the types and levels of code mixing. In the types of code mixing, the highest type was intra-sentential of code mixing (88.8%) and the lowest type was involving a change of pronunciation (0%). While, in the levels of code mixing, word level (44.8%) becomes the dominant while word repetition (3%) and idiom (1.5%) were the lowest.

This research is about code mixing used by Indonesian Influencers on Instagram. This researchaims to find out the types of code mixing used by Indonesian Influencers on Instagram. Thisresearch focuses on types of code mixing used by Indonesian Influncers by Musyken (2000) theory.The data then were analyzed using qualitative method. The research used observation anddocumentation to find out the data. The finding showed that there were 25 data of code mixingfrom 5 Indonesian Influencers and the research used 5 data to analyzed. The research concludedthat there are three types of code mixing used by Indonesian Influencers. They are insertion,alternation, and congruent lexicalization.

Informal short texts on the web are rich in emotions as they often reflect unfiltered immediate reactions to breaking news events. The emotion density, however, stands in contrast to its poverty of linguistic contexts and features for emotion classification. This paper tackles that challenge by proposing orthographic features based on orthographic code mixing and code-switching for both non-ML and ML approaches. Our results show that orthographic features routinely outperform grammatical features for emotion classification for short texts in all approaches as expected. Orthographic features were also shown to make more significant contributions, especially in terms of precision and in formal texts when state of the art deep learning algorithms are applied. This result confirms the effectiveness of the orthographic change feature to the task of emotion classification. These results are argued to be applicable to all languages because of the common code-shifting in languages with non-Latin orthographies, and the use of non-letter symbols in all languages.

The research aims to find the classification types of code switching and code mixing in twivortiare 2 novel then describe the influencing using types of code switching and code mixing in the novel. Descriptive qualitative method are used as the methodology in this research which means classify types of code switching and code mixing every sentences clearly through the dialogue of the novel. The research focused on the types of code switching (inter-sentential, intra-sentential, and tag-switching) and types of code mixing (insertion, alternation, and Congruent Lexicalization) while combine with influencing of used it. The results of the research many utterance are most of utterance using code mixing – insertation and code switching – intra-setential.

This research aims to decribe the use of pandemic covid-19 register in communication in various online media. This study uses descriptive qualitative research methods with a sosiolinguistic approach. The technique used is listen, note and screenshot. Based on the result of the analysis that has been done, the data is found in the form, function and use of language in the use of the pandemic covid-19 register in communication in various online media. The form of the use of registers is classified into three parts, namely lingua, limited environment and open environment. The function of using registers is classified as instrumental, regulatory and representational functions. Whereas the use of register usage language is classified as code switching and code mixing.

Despite their pervasiveness, current text-based conversational agents (chatbots) are predominantly monolingual, while users are often multilingual. It is well-known that multilingual users mix languages while interacting with others, as well as in their interactions with computer systems (such as query formulation in text-/voice-based search interfaces and digital assistants). Linguists refer to this phenomenon as code-mixing or code-switching. Do multilingual users also prefer chatbots that can respond in a code-mixed language over those which cannot? In order to inform the design of chatbots for multilingual users, we conduct a mixed-method user-study (N=91) where we examine how conversational agents, that code-mix and reciprocate the users' mixing choices over multiple conversation turns, are evaluated and perceived by bilingual users. We design a human-in-the-loop chatbot with two different code-mixing policies -- (a) always code-mix irrespective of user behavior, and (b) nudge with subtle code-mixed cues and reciprocate only if the user, in turn, code-mixes. These two are contrasted with a monolingual chatbot that never code-mixed. Users are asked to interact with the bots, and provide ratings on perceived naturalness and personal preference. They are also asked open-ended questions around what they (dis)liked about the bots. Analysis of the chat logs, users' ratings, and qualitative responses reveal that multilingual users strongly prefer chatbots that can code-mix. We find that self-reported language proficiency is the strongest predictor of user preferences. Compared to the Always code-mix policy, Nudging emerges as a low-risk low-gain policy which is equally acceptable to all users. Nudging as a policy is further supported by the observation that users who rate the code-mixing bot higher typically tend to reciprocate the language mixing pattern of the bot. These findings present a first step towards developing conversational systems that are more human-like and engaging by virtue of adapting to the users' linguistic style.

This study aims to determine the use of code mixing in the video study of Ustazah Mumpuni Handayayekti by using the analysis of sociolinguistic theories. The results of this study are using mixed code in lecturing Ustadzah Mumpuni Handayayekti into the use of: (1) mixed code of Javanese word forms, (2) mixed code of Javanese clause form codes, (3) mixed code of Balinese word forms, and (4) mixed code phrase of Javanese language. In addition, mixing the code out on the use (1) mixed form codes for Arabic clauses, (2) mixed form codes for Arabic words, (3) mixed form codes for Arabic phrases, and (4) mixed form codes for Indonesian words. Mixing the code in and mixing the code out used by Ustadzah Mumpuni Handayayekti in his lecture, shows that he is a multilingual speaker and can use variations of language well in the right context.

This researcher aims to determine the types and dominant types of code mixing used by sixth semester students of English Language Education Study Program at Cokroaminoto Palopo University. This research used a qualitative method. The instrument of this research was documentation. The result of this research indicates that (1) there are 64 codes mixing which is used by sixth semester students. The students used 4 types of code mixing. Those are 30 codes mixing in the form of word (46.9%); 28 codes mixing in the form of phrase insertion (43.8%); 5 codes mixing in hybrid insertion form (7.8%); and there is only 1 code mixing in the form of a reduplication/repetition insertion (1.5%). (2) The dominant type of code mixing used by the students is in the form of word insertion.

Code-switching and code-mixing, two by-products of language contact situations, are widely used linguistic strategies, not only in informal conversational contexts but also in the informal media discourse. The reasons hitherto applicable for the mixing of codes seem non-relevant in some code-mixing situations today. This is true to certain segments of the Sri Lankan media context, too, in that the popular Sri Lankan commercial broadcasting media is accused by prescriptivist groups of making excessive use of code-mixing. The purpose of this paper is to synthesize and explore the seminal research articles available in the field of code alternation so as to better understand the Sri Lankan situation of code-switching and code-mixing, as practiced by its general bilingual populations and by the popular commercial media channels.

Due to increase in the availability of numerous languages in the Web, cross language information retrieval is one of the happening issues in the field of natural language processing and information retrieval. Nowadays, people are habituated to combine two or more language words during oral or written discourse. Speakers have also employed intermixing of different languages and scripts in digital media while querying, blogging and on social media platforms. The way of representing two different language words of an utterance in their native scripts is known as mixed scripting. In the present work, we attempted to translate mixed script queries of Kannada and English languages into monolingual queries. We proposed three approaches for translation by constructing bilingual dictionary, word embeddings and Google translate. The proposed method outperforms the conventional dictionary based approach, when word embeddings were combined with the translations learnt from Google Translate and Dictionary.

This article addresses language identification at the word level in Indian social media corpora taken from Facebook, Twitter and WhatsApp posts that exhibit code-mixing between English-Hindi, English-Bengali, as well as a blend of both language pairs. Code-mixing is a fusion of multiple languages previously mainly associated with spoken language, but which social media users also deploy when communicating in ways that tend to be rather casual. The coarse nature of code-mixed social media text makes language identification challenging. Here, the performance of deep learning on this task is compared to feature-based learning, with two Recursive Neural Network techniques, Long Short Term Memory (LSTM) and bidirectional LSTM, being contrasted to a Conditional Random Fields (CRF) classifier. The results show the deep learners outscoring the CRF, with the bidirectional LSTM demonstrating the best language identification performance.

This research aimed at elaborating the variety of code-switching and code-mixing used in radio broadcast in East Kalimantan. This research used qualitative descriptive method and applied sociolinguistics study. In the analysis, the researcher employed Poplack and Muysken theory. Data sources were collected from radio broadcast in three radio stations representing three largest cities in East Kalimantan, such as Samarinda, Balikpapan, and Tenggarong. The type of data were conversation that appeared within the Pilihan Pendengar program from each of three stations: Gemaya 104.5 FM (Balikpapan), RB Radio 87.7 FM (Samarinda) and SBK 107.3 FM (Tenggarong). The research finds that there are varieties of code-switching, including: intra-sentential (40 data), inter-sentential (19 data) and extra-sentential (9 data). For code-mixing, this research finds three variants, which are alternation (24 data), insertion (14 data) and congruent lexicalization (8 data). The data are obtained from different languages, such as Indonesian, Javanese, Kutainese, English and Arabic.

The aim of the paper is to introduce the author's observations on the influence that English has on the Bulgarian forum "Netspeak" for women living in the USA. The collected data for the investigation are excerpted from a Bulgarian forum platform. The emphasis is on code-switching, one of the main topics in sociolinguistics; as well as on the interlanguage, which is a result of the penetration of some elements from English into the first language of communicators. The orthography may also be influenced as in Bulgaria the Cyrillic alphabet is used. The informal situation is a prerequisite for the mixed style use, especially in vocabulary. Different topics have been observed and discussed by users in the forum: the choice of baby accessories; baggage loss in airports; child diseases, child physical and mental development, etc. The investigated forum discussion includes 52,020 lexical items. Excerpted English words are few - only 1.5%. Most of them are nouns because of their easy grammatical adaptation. The verb group is not affected so noticeably. The Bulgarian verbal system is very complicated and English verbs used in Bulgarian context are grammatically adapted. There are some exceptions when people use ready-made constructions like idioms and phrasal verbs.

—This study examines code-switching and code mixing realized in the interactional activities undertaken by govermental leaders in the Denpasar City of Bali Province, Indonesia. Typically I reveal the language types used by bilingual and multingual governments during code swithing and code mixing events as well as investigating their leading factors. The study was designed using qualitative research method. The results of this study indicate that code mixing dan code-switching undertaken by the leaders in the governmental area of Denpasar city involves 3 types of language, namely Balinese, Indonesian, and English. An element of code mixing filler occurring during interactional process of the leaders is tangible words and phrases. Leading factors to such phenomena of code mixing and code-switching are: language ability, speech topics, speech participants, speech situations, prestige; and creating regional or religious nuances. Keywords: Code mixing; code switching; governmental leader

The NLP community has witnessed steep progress in a variety of tasks across the realms of monolingual and multilingual language processing recently. These successes, in conjunction with the proliferating mixed language interactions on social media have boosted interest in modeling code-mixed texts. In this work, we present CodemixedNLP, an open-source library with the goals of bringing together the advances in code-mixed NLP and opening it up to a wider machine learning community. The library consists of tools to develop and benchmark versatile model architectures that are tailored for mixed texts, methods to expand training sets, techniques to quantify mixing styles, and fine-tuned state-of-the-art models for 7 tasks in Hinglish. We believe this work has a potential to foster a distributed yet collaborative and sustainable ecosystem in an otherwise dispersed space of code-mixing research. The toolkit is designed to be simple, easily extensible, and resourceful to both researchers as well as practitioners., Comment: Accepted at the Fifth Workshop on Computational Approaches to Linguistic Code-Switching-CALCS 2021

Virtual Assistants like Bixby, Google Assistant, Cortana & Alexa are making life easier by understanding intent from user's input utterance. Adoption of Virtual Assistants in multilingual society like India is harder than in monolingual communities. People in multilingual societies tend to use linguistic elements from multiple languages in a single sentence. This phenomenon produces code-mixing, which lacks standard syntactic & semantic linguistic properties unlike monolingual input. It is a challenging NLP Task to understand user's intent from code-mixed input. There is a lack of relevant corpus & researches for Intent Classification on Code-mixed text. In this paper, we introduce first of its kind a Hindi-English Code-Mixed dataset for Intent Classification (CoMTIC), covering 10 most preferred features for a Virtual Assistant. We also introduce a novel deep learning based method of Intent classification from Code-Mixed Text. We conduct empirical analysis by comparing the suitability & performance with various state-of-the-art methods. Our method attains 96.68% accuracy on our dataset.

We describe models focused at the understudied problem of translating between monolingual and code-mixed language pairs. More specifically, we offer a wide range of models that convert monolingual English text into Hinglish (code-mixed Hindi and English). Given the recent success of pretrained language models, we also test the utility of two recent Transformer-based encoder-decoder models (i.e., mT5 and mBART) on the task finding both to work well. Given the paucity of training data for code-mixing, we also propose a dependency-free method for generating code-mixed texts from bilingual distributed representations that we exploit for improving language model performance. In particular, armed with this additional data, we adopt a curriculum learning approach where we first finetune the language models on synthetic data then on gold code-mixed data. We find that, although simple, our synthetic code-mixing method is competitive with (and in some cases is even superior to) several standard methods (backtranslation, method based on equivalence constraint theory) under a diverse set of conditions. Our work shows that the mT5 model, finetuned following the curriculum learning procedure, achieves best translation performance (12.67 BLEU). Our models place first in the overall ranking of the English-Hinglish official shared task., Comment: Computational Approaches to Linguistic Code-Switching (CALCS 2021) workshop

The use of language (function of language) according to Halliday (1985) is a language used as a means of communication in society. The word usage can simply be said to have the same meaning as the function of language. The purpose of this study was to determine the style of language used in the tourism industry. when talking about the use of language, it means the way someone uses their language or their languages if they know more than one language. Why these locations were chosen as the research locations because these areas were considered as bilingual communities, so they were very suitable and in accordance with the objectives of this study. The choice of locations in the tourism areas of South Bali due to the fact that these tourism destinations are the center of tourism in Badung and the speaking community mostly comes from various regions in Indonesia and abroad. Language is a communication tool used by humans, including speakers in the tourism industry in their interactions with guests, fellow employees, as well as superiors and subordinates to convey intentions, thoughts and feelings. The languages used by speakers in the tourism area of South Bali include Indonesian, Balinese, foreign languages such as English, French, Japanese and others. To achieve these communication objectives the speakers used communication strategies including, code switching, code mixing and also the use of jargon related to the tourism industry.

The study focuses on discussing the types of English-Indonesian switching and mixing and the factor causing of English- Indonesian code switching and code mixing. All of data were taken from the Skinnyindonesian24 YouTube channel from two brothers of Da Lopez they are Jovial da Lopez or Kajo and Andovi da Lopez. The data is taken from School Tour video released on October 30, 2018. The data collection method in this study is library research. The library research is done by watching the School Tour video by Skinnyindonesian24 repeatedly. There are three steps in data collection. First, watch the School Tour video carefully and repeatedly to understand the content. The data were analyzed using the qualitative method, the theory applied in this study were based on the theory proposed by Wardhaugh. The result of the study shown that there were three types of code switching and code mixing found, they are tag switching inter-sentential switching, intra-switching and insertion, alternation, congruent lexicalization. The factors that caused the writer to used code switching are four factors and all of them are contained in video the Skinnyindonesian24. There are also several factors that cause the author to use code mixing, from five factors, there were three factors are found in the data, they are language dominance, language attitude, psycholinguistic motivation.

Multilingual models have demonstrated impressive cross-lingual transfer performance. However, test sets like XNLI are monolingual at the example level. In multilingual communities, it is common for polyglots to code-mix when conversing with each other. Inspired by this phenomenon, we present two strong black-box adversarial attacks (one word-level, one phrase-level) for multilingual models that push their ability to handle code-mixed sentences to the limit. The former uses bilingual dictionaries to propose perturbations and translations of the clean example for sense disambiguation. The latter directly aligns the clean example with its translations before extracting phrases as perturbations. Our phrase-level attack has a success rate of 89.75% against XLM-R-large, bringing its average accuracy of 79.85 down to 8.18 on XNLI. Finally, we propose an efficient adversarial training scheme that trains in the same number of steps as the original model and show that it improves model accuracy., Comment: To be presented at NAACL-HLT 2021. Abstract also published in the Rising Stars Track of the Workshop on Computational Approaches to Linguistic Code-Switching (CALCS 2021)

Sentiment analysis of Dravidian languages has received attention in recent years. However, most social media text is code-mixed and there is no research available on sentiment analysis of code-mixed Dravidian languages. The Dravidian-CodeMix-FIRE 2020, a track on Sentiment Analysis for Dravidian Languages in Code-Mixed Text, focused on creating a platform for researchers to come together and investigate the problem. There were two languages for this track: (i) Tamil, and (ii) Malayalam. The participants were given a dataset of YouTube comments and the goal of the shared task submissions was to recognise the sentiment of each comment by classifying them into positive, negative, neutral, mixed-feeling classes or by recognising whether the comment is not in the intended language. The performance of the systems was evaluated by weighted-F1 score.

Usage of social media has become more widespread to express sentiment, emotion about public events, government policies, product reviews etc. Performing Sentiment Analysis (SA) on social media data will give more and more insights about user’s behavior. Multilingual society like India, it is very common to use code switch text in social media to express their views. Switching between languages while communicating is refer as code mixing or code switching. Analyzing this code switch text and getting the useful information from this too harder because of its unstructured linguistic nature. In this paper, we proposed a hybrid model called SAEKCS for sentiment analysis on Kannada-English code switch text. Our proposed model uses deep learning techniques like Convolutional Neural Network (CNN) and Bidirectional Long Short Term Memory (BiLSTM) for sentiment analysis in code switch text. Our experimental results shows that 77.6% of accuracy and 69.6% of coverage. These results are much better than existing works [17] [18].

Code mixing is a common phenomena in multilingual societies where people switch from one language to another for various reasons. Recent advances in public communication over different social media sites have led to an increase in the frequency of code-mixed usage in written language. In this paper, we present the Generative Morphemes with Attention (GenMA) Model sentiment analysis system contributed to SemEval 2020 Task 9 SentiMix. The system aims to predict the sentiments of the given English-Hindi code-mixed tweets without using word-level language tags instead inferring this automatically using a morphological model. The system is based on a novel deep neural network (DNN) architecture, which has outperformed the baseline F1-score on the test data-set as well as the validation data-set. Our results can be found under the user name "koustava" on the "Sentimix Hindi English" page, To be published in 14th International Workshop on Semantic Evaluation SemEval-2020

Code mixing (or code switching) is a common phenomenon observed in social-media content generated by a linguistically diverse user-base. Studies show that in the Indian sub-continent, a substantial fraction of social media posts exhibit code switching. While the difficulties posed by code mixed documents to further downstream analyses are well-understood, lending visibility to code mixed documents under certain scenarios may have utility that has been previously overlooked. For instance, a document written in a mixture of multiple languages can be partially accessible to a wider audience; this could be particularly useful if a considerable fraction of the audience lacks fluency in one of the component languages. In this paper, we provide a systematic approach to sample code mixed documents leveraging a polyglot embedding based method that requires minimal supervision. In the context of the 2019 India-Pakistan conflict triggered by the Pulwama terror attack, we demonstrate an untapped potential of harnessing code mixing for human well-being: starting from an existing hostility diffusing hope speech classifier solely trained on English documents, code mixed documents are utilized to perform cross-lingual sampling and retrieve hope speech content written in a low-resource but widely used language - Romanized Hindi. Our proposed pipeline requires minimal supervision and holds promise in substantially reducing web moderation efforts. A further exploratory study on a new COVID-19 data set introduced in this paper demonstrates the generalizability of our cross-lingual sampling technique.

[EN] India is a nation of geographical and cultural diversity where over 1600 dialects are spoken by the people. With the technological advancement, penetration of the internet and cheaper access to mobile data, India has recently seen a sudden growth of internet users. These Indian internet users generate contents either in English or in other vernacular Indian languages. To develop technological solutions for the contents generated by the Indian users using the Indian languages, the Forum for Information Retrieval Evaluation (FIRE) was established and held for the first time in 2008. Although Indian languages are written using indigenous scripts, often websites and user-generated content (such as tweets and blogs) in these Indian languages are written using Roman script due to various socio-cultural and technological reasons. A challenge that search engines face while processing transliterated queries and documents is that of extensive spelling variation. MSIR track was first introduced in 2013 at FIRE and the aim of MSIR was to systematically formalize several research problems that one must solve to tackle the code mixing in Web search for users of many languages around the world, develop related data sets, test benches and most importantly, build a research community focusing on this important problem that has received very little attention. This document is a comprehensive report on the 4 years of MSIR track evaluated at FIRE between 2013 and 2016., Somnath Banerjee and Sudip Kumar Naskar are supported by Media Lab Asia, MeitY, Government of India, under the Visvesvaraya PhD Scheme for Electronics & IT. The work of Paolo Rosso was partially supported by the MISMIS research project PGC2018-096212-B-C31 funded by the Spanish MICINN.

Simplified Chinese to Traditional Chinese character conversion is a common preprocessing step in Chinese NLP. Despite this, current approaches have poor performance because they do not take into account that a simplified Chinese character can correspond to multiple traditional characters. Here, we propose a model that can disambiguate between mappings and convert between the two scripts. The model is based on subword segmentation, two language models, as well as a method for mapping between subword sequences. We further construct benchmark datasets for topic classification and script conversion. Our proposed method outperforms previous Chinese Character conversion approaches by 6 points in accuracy. These results are further confirmed in a downstream application, where 2kenize is used to convert pretraining dataset for topic classification. An error analysis reveals that our method's particular strengths are in dealing with code-mixing and named entities., Comment: Accepted to ACL 2020

Code switching is a linguistic phenomenon which may occur within a multilingual setting where speakers share more than one language. With the increasing communication between groups with different languages, this phenomenon is more and more popular. However, there are little research and data in this area, especially in code-mixing sentiment classification. In this work, the domain transfer learning from state-of-the-art uni-language model ERNIE is tested on the code-mixing dataset, and surprisingly, a strong baseline is achieved. And further more, the adversarial training with a multi-lingual model is used to achieved 1st place of SemEval-2020 Task9 Hindi-English sentiment classification competition.

In this paper, we present an approach for sentiment analysis in code-mixed language on twitter defined in SemEval-2020 Task 9. Our team (referred as LiangZhao) employ different multilingual models with weighted loss focused on complexity of code-mixing in sentence, in which the best model achieved f1-score of 0.806 and ranked 1st of subtask- Sentimix Spanglish. The performance of method is analyzed and each component of our architecture is demonstrated.

In this paper, we present our system for the SemEval 2020 task on code-mixed sentiment analysis. Our system makes use of large transformer based multilingual embeddings like mBERT. Recent work has shown that these models posses the ability to solve code-mixed tasks in addition to their originally demonstrated cross-lingual abilities. We evaluate the stock versions of these models for the sentiment analysis task and also show that their performance can be improved by using unlabelled code-mixed data. Our submission (username Genius1237) achieved the second rank on the English-Hindi subtask with an F1 score of 0.726.

The use of multilingualism in the new generation is widespread in the form of code-mixed data on social media, and therefore a robust translation system is required for catering to the monolingual users, as well as for easier comprehension by language processing models. In this work, we present a translation framework that uses a translation-transliteration strategy for translating code-mixed data into their equivalent monolingual instances. For converting the output to a more fluent form, it is reordered using a target language model. The most important advantage of the proposed framework is that it does not require a code-mixed to monolingual parallel corpus at any point. On testing the framework, it achieved BLEU and TER scores of 16.47 and 55.45, respectively. Since the proposed framework deals with various sub-modules, we dive deeper into the importance of each of them, analyze the errors and finally, discuss some improvement strategies., Comment: 6 pages, 3 figures, 2 tables

Language identification of social media text still remains a challenging task due to properties like code-mixing and inconsistent phonetic transliterations. In this paper, we present a supervised learning approach for language identification at the word level of low resource Bengali-English code-mixed data taken from social media. We employ two methods of word encoding, namely character based and root phone based to train our deep LSTM models. Utilizing these two models we created two ensemble models using stacking and threshold technique which gave 91.78% and 92.35% accuracies respectively on our testing data., 6 pages, 5 figures, 5 tables

With internet and social media breaking the ice, more and more people across the globe have started to use social media platforms like Facebook, Twitter, Instagram etc. Most people follow Multilingualism as a mode of communication to convey information across the globe. They share topics over the common forum to converse, with the use of multiple languages being spoken either by individual speaker or group of speakers. This essentially makes the context more complex to understand and it makes even more harder for processing various Natural Language Processing (NLP) tasks. Such user behavior of mixing multiple languages in one single discussion topic, having multiple community inclusion is referred as code-switching. At IEEE 2019 Big data conference, a Shared Task (Understanding Multilingual Communities through Analysis of code-switching Behaviors in Social Media Discussions) is conducted as a track of Big Data Cup. Firstly, Tasks is to detect the language of each post given in the discussion forum with the help of multiple languages. Secondly, to detect relevance score of a post by determining how much the content is closely connected or appropriate in the discussion. This paper proposes a novel approach to detect the language of each word in the post using Natural Language Processing (NLP) techniques involving linguistics, Python package(langdetect) and various other approaches. It also explains how Machine Learning is applied to figure out relevance of a post and other metrics required for prediction. Code-Mixing detection is an important step for any NLP application to determine the language of a post at first place in order to perform any NLP task over social media.

This research aimed to explore the cultural wisdom which encouraged the spirit to empower the community and to reveal code-mixing utilization in the conversation. This research was conducted in progressive farming communities in West Java and Banten provinces. They communicated through an ODF (Online Discussion Forum) bridged by WhatsApp Group (WAG) social networking app. Content Analysis method could reveal the wisdom of the Sundanese culture that underlined the spirit of the progressive farmers. The observation period was from August 25 - November 26, 2016. During the observation period, there were 10.313 statements emerged in the discussion. Nvivo 8 software was used to analyze discussion themes. 20 statements were displayed in the results section as they could reflect the research questions. In code-mixing Indonesian and Sundanese, the progressive farmers also expressed their efforts to nurture young farmers by conducting training. The result shows that code-mixing of Indonesian and local language as mother tongue in the discussions accelerates mutual understanding because of cultural similarity between actors. It also implies that communication technology could be utilized for maintaining indigenous cultural wisdom.

Abuse in cyber space is a problem requiring immediate attention. Unfortunately, despite advances in Natural Language Processing techniques, there are clear limitations in detecting instances of cyber abuse today. Challenges arising due to different languages that teens communicate with today, and usage of codes along with code mixing and code switching make the design of a comprehensive approach very hard. Existing NLP based approaches for detecting cyber abuse thus suffer from a high degree of false negatives and positives. In this paper, we investigate a new approach to detect instances of cyber abuse. Our approach is motivated by the premise that abusers tend to have unique facial expressions while engaging in an actual abuse episode, and if we are successful, such an approach will be language-agnostic. Here, using only four carefully identified facial features without any language processing, and realistic experiments with 15 users, our system proposed in this paper achieves 98% accuracy for same-user evaluation and up to 74% accuracy for cross-user evaluation in detecting instances of cyber abuse.

Nowadays, English words mixed in Thai speech are usually found in a typical speaking style. Consequently, to increase the performance of the speech recognition system, a Thai-English code-mixing speech corpus is required. This paper describes the design and construction of LOTUS-BI corpus: a Thai-English code-mixing speech corpus aimed to be the essential speech database for training acoustic model and language model in order to obtain the better speech recognition accuracy. LOTUS-BI corpus contains 16.5 speech hours from 4 speech tasks: interview, talk, seminar, and meeting. Now, 11.5 speech hours of data from the interview, talk, and seminar acquire from the internet have been transcribed and annotated. Whereas, the rest of 5 speech hours from meeting task has been transcribing. Therefore, only 11.5 speech hours of data were analyzed in this paper. Furthermore, the pronunciation dictionary of vocabularies from LOTUS-BI corpus is created based on Thai phoneme set. The statistical analysis of LOTUS-BI corpus revealed that there are 37.96% of code-mixing utterances, including 34.23% intra-sentential and 3.73% inter-sentential utterances. The occurrence of English vocabularies is 29.04% of the total vocabularies in the corpus. Besides, nouns are found in 90% of all English vocabularies in the corpus and 10% in the other grammatical categories.

[EN] Before the advent of the Internet era, code-mixing was mainly used in the spoken form. However, with the recent popular informal networking platforms such as Facebook, Twitter, Instagram, etc., in social media, code-mixing is being used more and more in written form. User-generated social media content is becoming an increasingly important resource in applied linguistics. Recent trends in social media usage have led to a proliferation of studies on social media content. Multilingual social media users often write native language content in non-native script (cross-script). Recently Banerjee et al. [9] introduced the code-mixed cross-script question answering research problem and reported that the ever increasing social media content could serve as a potential digital resource for less-computerized languages to build question answering systems. Question classification is a core task in question answering in which questions are assigned a class or a number of classes which denote the expected answer type(s). In this research work, we address the question classification task as part of the code-mixed cross-script question answering research problem. We combine deep learning framework with feature engineering to address the question classification task and enhance the state-of-the-art question classification accuracy by over 4% for code-mixed cross-script questions., The work of the third author was partially supported by the SomEMBED TIN2015-71147-C2-1-P MINECO research project.

This research deals with code mixing in kiss 105 Fm Medan. The aims are to find; (1) the type of code mixing in kiss 105 Fm Medan (2) how the announcers mix language in the process of broadcasting. Code mixing occurs when speakers mix and insert foreign words (other code) in the dominant language used (Bloomfield, 1972). The whole process of analysis is conducted with descriptive qualitative method focusing on a detailed examination of one setting or a single subject, a single depository of document, or one particular event. (Moleong, 2000). The data are from the utterances of announcers which contain code mixing during the process of broadcasting in two different radio programs; kiss weekly hot forty and movie buzz. The data are analyzed by using some steps as normally applied in descriptive qualitative method. The results show that the three types of code mixing, namely intra-sentential, intra-lexical, and a change of pronunciation are found in the utterances of the announcers. The intra sentential is dominantly found in this research. The total number of intra sentential is 38, while the total number of intra lexical is 6.