Linguistic Data Consortium Research Articles

An Automatic Speaker Clustering Pipeline for the Air Traffic Communication Domain

In air traffic management (ATM), voice communications are critical for ensuring the safe and efficient operation of aircraft. The pertinent voice communications—air traffic controller (ATCo) and pilot—are usually transmitted in a single channel, which poses a challenge when developing automatic systems for air traffic management. Speaker clustering is one of the challenges when applying speech processing algorithms to identify and group the same speaker among different speakers. We propose a pipeline that deploys (i) speech activity detection (SAD) to identify speech segments, (ii) an automatic speech recognition system to generate the text for audio segments, (iii) text-based speaker role classification to detect the role of the speaker—ATCo or pilot in our case—and (iv) unsupervised speaker clustering to create a cluster of each individual pilot speaker from the obtained speech utterances. The speech segments obtained by SAD are input into an automatic speech recognition (ASR) engine to generate the automatic English transcripts. The speaker role classification system takes the transcript as input and uses it to determine whether the speech was from the ATCo or the pilot. As the main goal of this project is to group the speakers in pilot communication, only pilot data acquired from the classification system is employed. We present a method for separating the speech parts of pilots into different clusters based on the speaker’s voice using agglomerative hierarchical clustering (AHC). The performance of the speaker role classification and speaker clustering is evaluated on two publicly available datasets: the ATCO2 corpus and the Linguistic Data Consortium Air Traffic Control Corpus (LDC-ATCC). Since the pilots’ real identities are unknown, the ground truth is generated based on logical hypotheses regarding the creation of each dataset, timing information, and the information extracted from associated callsigns. In the case of speaker clustering, the proposed algorithm achieves an accuracy of 70% on the LDC-ATCC dataset and 50% on the more noisy ATCO2 dataset.

A Language Model Optimization Method for Turkish Automatic Speech Recognition System

The current Automatic Speech Recognition (ASR) modeling strategy still suffers from huge performance degradation when faced with languages with limited resources such as Turkish. Especially when the Language Model (LM) does not support the Acoustic Model (AM) sufficiently, the Word Error Rate (WER) increases. Therefore, a robust LM makes a strong contribution to improving ASR performance by generating word relations from the existing corpus. However, developing a robust language model is a challenging task due to the agglutinative nature of Turkish. Therefore, within the scope of the study, a sentence-level LM optimization method is proposed to improve the WER performance of Turkish ASR. In the proposed method, instead of a fixed word sequence obtained from the Markov assumptions, the probability of the word sequence forming a sentence was calculated. A method with n-gram and skip-gram properties is presented to obtain the word sequence probability. The proposed method has been tested on both statistical and Artificial Neural Network (ANN) based LMs. In the experiments carried out using, not only words but also sub-word level, two Turkish corpora (METU and Bogazici) shared via Linguistic Data Consortium (LDC) and a separate corpus, which we separate corpus that we specially created as HS was used. According to the experimental results obtained from statistical-based LM, 0.5% WER increases for the METU corpus, 1.6% WER decreases for the Bogazici corpus, and a 2.5% WER decrease for the HS corpus were observed. In the Feedforward Neural Networks (FNN) based LM, WER decreases were observed 0.2% for the METU corpus, 0.8% for the Bogazici corpus, and 1.6% for the HS corpus. Also, in the Recurrent Neural Network (RNN)-Long Short Term Memory (LSTM) based LM, WER decreases were observed 0.6% for METU corpus, 1.1% for the Bogazici corpus and 1.5% for the HS corpus. As a result, when the proposed method was applied to the LMs required for ASR, WER decreased, and the total performance of ASR increased.

Linguistic Data Consortium

The Linguistic Data Consortium (LDC) creates, stores, and distributes language resources, in particular language corpora, for educational, governmental, and commercial purposes. At this time, the number of corpora available on the LDC platform exceeds 900, and new ones are added on a monthly basis. Since its founding in 1992, LDC has had thousands of members such as universities, libraries, government research laboratories, and corporations. While there is a plethora of language resources available to researchers these days, LDC continues to stand out by offering an unparalleled selection of carefully curated corpora combined with a variety of linguistic tools.

Text-Independent Automatic Dialect Recognition of Marathi Language using Spectro-Temporal Characteristics of Voice

Text-independent dialect recognition system is proposed in this paper for Marathi language. India is rich in language varieties. Each language in turn has its unique dialect variations. Maharashtra has Marathi as official language and for Goa it is a co-official language . In literature there are very few studies available for Indian language recognition and then their respective dialect recognition. So research work available for regional languages such as Marathi is extremely limited. As a part of research work, an attempt is made to generate a case study of a low resourced Marathi language dialect recognition system. The study was carried out using Marathi speech data corpus provided by Linguistic Data Consortium for Indian Language (LDC- IL). This corpus includes four major dialects of Marathi speakers. The efficiency and performance evaluation of the explored spectral (rhythmic) and temporal features are carried out to perform classification tasks. We investigated the performance of six different classifiers; K-nearest neighbor (KNN), Naïve Bayes (NB), Support Vector Machine (SVM), Decision Tree (DT) classifier , Stochastic Gradient Descent (SGD) classifier and Ridge Classifier (RC). Experimental results have demonstrated that the RC classifier worked well with 84.24% of accuracy for fifteen spectral and temporal features. With twelve MFCCs it has been observed that SGD has outperformed among all classifiers with accuracy of 80.63%. For further study, a prominent feature subset as a part of dimensionality reduction has been identified using chi square, mutual information and ANOVA-f test. In this chi-square based feature extraction method has proven to be the best over over mutual information and ANOVA f-test.

Towards developing speaker diarization for parent-child interactions

Daily interactions of children with their parents are crucial for spoken language skills and overall development. Capturing such interactions can help to provide meaningful feedback to parents as well as practitioners. Naturalistic audio capture and developing further speech processing pipeline for parent-child interactions is a challenging problem. One of the first important steps in the speech processing pipeline is Speaker Diarization—to identify who spoke when. Speaker Diarization is the method of separating a captured audio stream into analogous segments that are differentiated by the speaker’s (child or parent’s) identity. Following ongoing COVID-19 restrictions and human subjects research IRB protocols, an unsupervised data collection approach was formulated to collect parent-child interactions (of consented families) using LENA device—a light weight audio recorder. Different interaction scenarios were explored: book reading activity at home and spontaneous interactions in a science museum. To identify child’s speech from a parent, we train the Diarization models on open-source adult speech data and children speech data acquired from LDC (Linguistic Data Consortium). Various speaker embeddings (e.g., x-vectors, i-vectors, resnets) will be explored. Results will be reported using Diarization Error Rate. [Work sponsored by NSF via Grant Nos. 1918032 and 1918012.]

Signalling conditional relations

Abstract We investigate how discourse relations and their subtypes are signalled, extending the set of discourse signals from connectives and lexical cue phrases to the wide range of semantic, syntactic, and orthographic signals of the RST Signalling Corpus (Das, Debopam & Maite Taboada. 2018. RST signalling corpus. Language Resources and Evaluation 52. 149–184). This extension requires re-evaluating previous predictions on discourse signalling, in particular, those of Sanders, Ted. 2005. Coherence, causality and cognitive complexity in discourse. In M. Aurnague, M. Bras, A. Le Draoulec & L. Vieu (eds.), Proceedings/Actes SEM-05, first international symposium on the exploration and modelling of meaning, 105–114. Biarritz causality-by-default hypothesis, the hypothesis of uniform information density (Frank, Austin & Florian Jaeger. 2008. Speaking rationally: Uniform information density as an optimal strategy for language production. In Proceedings of the 30th annual meeting of the Cognitive Science Society, 933–938. https://escholarship.org/uc/item/7d08h6j4 (accessed 18 May 2022)), and the hypothesis that discourse is continuous by preference (Segal, Erwin, Judith Duchan & Paula Scott. 1991. The role of interclausal connectives in narrative structuring. Discourse Processes 14. 27–54; Murray, John. 1997. Connectives and narrative text. Memory and Cognition 25. 227–236). We evaluate the predictions of these theories on the conditional relations in the RST Discourse Treebank (Carlson, Lynn, Daniel Marcu & Mary Ellen Okurowski. 2002. RST Discourse Treebank. LDC2002T07. Philadelphia: Linguistic Data Consortium), using causal relations as a control group. Informativity and continuity are operationalized in terms of semantic complexity and Givón, Talmy. 1993. English grammar: A function-based introduction, vol. 2. Amsterdam: John Benjamins dimensions of deictic shift. Our results show that the hypotheses make accurate predictions only for the relation groups in their entirety but not for the observed in-group variation, in particular, the low amount of marking for the hypothetical subtype of conditional relations. We attribute this difference to the distribution of intra- and inter-sentential occurrences across the conditional subtypes: intra-sentential relations are consistently more marked than inter-sentential ones, and hypothetical relations are special in that they appear predominantly inter-sententially.

Research on an English translation method based on an improved transformer model

Abstract With the expansion of people’s needs, the translation performance of traditional models is increasingly unable to meet current demands. This article mainly studied the Transformer model. First, the structure and principle of the Transformer model were briefly introduced. Then, the model was improved by a generative adversarial network (GAN) to improve the translation effect of the model. Finally, experiments were carried out on the linguistic data consortium (LDC) dataset. It was found that the average Bilingual Evaluation Understudy (BLEU) value of the improved Transformer model improved by 0.49, and the average perplexity value reduced by 10.06 compared with the Transformer model, but the computation speed was not greatly affected. The translation results of the two example sentences showed that the translation of the improved Transformer model was closer to the results of human translation. The experimental results verify that the improved Transformer model can improve the translation quality and be further promoted and applied in practice to further improve the English translation and meet application needs in real life.

G-Cocktail: An Algorithm to Address Cocktail Party Problem of Gujarati Language Using Cat Boost

This paper is an attempt to address to the problem of native language in a mixed voice environment. G- Cocktail would aid these applications in identifying commands given in Gujarati, even from a mixed voice stream. There are two phases of G-cocktail in the first phase, it creates features after filtering the voices and in the second it trains and classifies the dataset. This trained dataset helps in recognizing the new voice signal. The challenge in training a native language is the availability of a small dataset. A single-word input is used in model and phrase benchmark dataset from Microsoft and the Linguistic Data Consortium for Indian Languages (LDC-IL). To overcome the over fitting problem due to smaller dataset we used CatBoost algorithm. And fine-tuned the classification model to avoid the over fitting issue. Mel-frequency cepstral coefficients (MFCCs) are coefficients that collectively make up an MFC. They are derived from a type of cepstral representation of the audio clip (a nonlinear "spectrum-of-a-spectrum"). MFCC is good for human voices but noises in the sound makes it less productive. To avoid this shortcoming of MFCC, first filtered the voices are used and then calculated the MFCC. The most relevant features are retained to make it more robust. With MFCC features, the pitch of the voices is also added, as pitch could vary with regional changes, mood of the person, age, and knowledge of the language to the speaker. A voice print of the whole sound files is constructed and fed it as features to the classification model. For training and testing 70% and 30% ratio is used in algorithms like K-means, Naïve Bayes, and Light GBM. Proposed model is compared with given data set and results proved that G-cocktail using XBoost performed better than the others under the given scenario in all parameters.

King Saud University Emotions Corpus: Construction, Analysis, Evaluation, and Comparison

Emotional speech recognition for the Arabic language is insufficiently tackled in the literature compared to other languages. In this paper, we present the work of creating and verifying the King Saud University Emotions (KSUEmotions) corpus, which was released by the Linguistic Data Consortium (LDC) in 2017 as the first public Arabic emotional speech corpus. KSUEmotions contains an emotional speech of twenty-three speakers from Saudi Arabia, Syria, and Yemen, and includes the emotions: neutral, happiness, sadness, surprise, and anger. The corpus content is verified in two different ways: a human perceptual test by nine listeners who rate emotional performance in audio files, and automatic emotion recognition. Two automatic emotion recognition systems are experimented with: Residual Neural Network and Convolutional Neural Network. This work also experiments with emotion recognition for the English language using the Emotional Prosody Speech and Transcripts Corpus (EPST). The current experimental work is conducted in three tracks: (i) monolingual, where independent experiments for Arabic and English are carried out, (ii) multilingual, where the Arabic and English corpora are merged in as mixed corpus, and (iii) cross-lingual, where models are trained using one language and tested using the other. A challenge encountered in this work is that the two corpora do not contain the same emotions. That problem is tackled by mapping the emotions to the arousal-valance space.

Arabic speaker recognition system based on phoneme fusion

With the increasing number and popularity of smart devices in the past few decades, especially those that execute different types of health-related applications, their security is a growing concern. Biometric authentication can be used for device security, and speaker recognition (SR) is one of its elegant forms. In this study, an SR system is developed for the security of smart devices, which identifies a person from the shortest utterance of Arabic language, achieving the highest SR rate, by using the benchmark Linguistic Data Consortium (LDC) Arabic dataset. This study focuses on the use of the consonants and vowels of the Arabic language. It is observed that certain consonants or a fusion of certain consonants in Arabic language helps to ascertain the speaker’s identity efficiently. The shortest utterance that better presents the speaker’s identity by using such consonants is revealed. The system was analyzed by using different numbers of consonants and achieved the maximum (100%) SR rate using a fusion of only three consonants of the Arabic language. These consonants or their combinations can be used to develop text for any speaker-based authentication system.

Building a Speech and Text Corpus of Turkish: Large Corpus Collection with Initial Speech Recognition Results

To build automatic speech recognition (ASR) systems with a low word error rate (WER), a large speech and text corpus is needed. Corpus preparation is the first step required for developing an ASR system for a language with few argument speech documents available. Turkish is a language with limited resources for ASR. Therefore, development of a symmetric Turkish transcribed speech corpus according to the high resources languages corpora is crucial for improving and promoting Turkish speech recognition activities. In this study, we constructed a viable alternative to classical transcribed corpus preparation techniques for collecting Turkish speech data. In the presented approach, three different methods were used. In the first step, subtitles, which are mainly supplied for people with hearing difficulties, were used as transcriptions for the speech utterances obtained from movies. In the second step, data were collected via a mobile application. In the third step, a transfer learning approach to the Grand National Assembly of Turkey session records (videotext) was used. We also provide the initial speech recognition results of artificial neural network and Gaussian mixture-model-based acoustic models for Turkish. For training models, the newly collected corpus and other existing corpora published by the Linguistic Data Consortium were used. In light of the test results of the other existing corpora, the current study showed the relative contribution of corpus variability in a symmetric speech recognition task. The decrease in WER after including the new corpus was more evident with increased verified data size, compensating for the status of Turkish as a low resource language. For further studies, the importance of the corpus and language model in the success of the Turkish ASR system is shown.

The Stance and Subjectivity in the Use of the Korean Post-Verbal Negation '-ci anh-' in the Telephone Conversation

조재현. 2018. 전화 대화에 나타난 한국어 장형 부정 ‘-지 않-’의 스탠스와 주관성 분석. 국제한국어교육 4(2), 81-106. 본 연구는 실제 전화 대화상에서 쓰인 한국어의 장형 부정 표현인 ‘-지 않-’이 드러내는 스탠스와 주관성(subjectivity)을 코퍼스 분석을 통해 살펴본 연구이다. 한국어의 장형과 단형 부정 표현은 전통적으로 통사론의 영역에서 다루 어져 왔는데 주로 그 둘 간의 교체가능성과 이들 각각이 가지는 부정의 범위가 논의의 초점이 되어 왔다. 최근에는 구어의 영역에서 이들 표현의 상호작용적 기능을 밝히고자 하는 연구가 활발하게 진행되어 왔는 데, 기존 연구의 성과에 기여하기 위해 본 연구는 ‘-지 않-’의 기능적, 상호작용적 특성을 아래와 같은 방법을 통해 조명해 보았다. 주관성의 표출과 관련하여 ‘-지 않-’과 빈번하게 함께 쓰이는 주어와 술어의 유형이 있는지, 자연스러운 전화 대화에서 장형 부정 표현이 특정한 스탠 스를 나타내는 수사적 용법으로 쓰이는 사용상의 패턴이 있는지, ‘-지 않-’과 높은 빈도로 함께 쓰이는 특정한 어미 표현이 있는지, 그리고 마지막으로 이들 표현이 대화 상의 어느 위치에 자주 나타나는지에 대한 분석이 이루어졌다. (캘리포니아주립대학교, 로스엔젤레스)Jaehyun Jo. 2018. The Stance and Subjectivity in the Use of the Korean Post-Verbal Negation -ci anh- in the Telephone Conversation. International Journal of Korean Language Education 4(2), 81-106. This corpus-based study demonstrates the results of analyses of the Korean negative structure ‘-ci anh-’ in terms of its subjectivity and stances in naturally occurring telephone conversations. The Korean post-verbal negation ‘-ci anh-’ and pre-verbal negation ‘an’ have been studied in syntax and semantics for its interchangeability and the scope of each negation type. Contributing to more recent studies that attempt to find the interactional functions of these negatives in spoken discourse, this study examines the functional and interactional characteristics of ‘-ci anh-’ structure appearing in the Linguistic Data Consortium (LDC) Call-Friend Korean corpus of 100 telephone calls. In the analyses, the utterances were coded with following information: (1) what subject and predicate types appear with ‘-ci anh-’ in relation to how its subjectivity is established; (2) whether each of ‘-ci anh-’ tokens is a simple negative or a rhetorical question (or statement) that shows the speaker’s specific stance; (3) also with what kind of ending combinations and in which position of one’s turn the post-verbal negation was deployed. (University of California, Los Angeles)

Improving the performance of the speaker emotion recognition based on low dimension prosody features vector

Speaker emotion recognition is an important research issue as it finds lots of applications in human–robot interaction, computer–human interaction, etc. This work deals with the recognition of emotion of the speaker from speech utterance. For that features like pitch, log energy, zero crossing rate, and first three formant frequencies are used. Feature vectors are constructed using the 11 statistical parameters of each feature. The Artificial Neural Network (ANN) is chosen as a classifier owing to its universal function approximation capabilities. In ANN based classifier, the time required for training the network as well as for classification depends upon the dimension of feature vector. This work focused on development of a speaker emotion recognition system using prosody features as well as reduction of dimensionality of feature vectors. Here, principle component analysis (PCA) is used for feature vector dimensionality reduction. Emotional prosody speech and transcription from Linguistic Data Consortium (LDC) and Berlin emotional databases are considered for evaluating the performance of proposed approach for seven types of emotion recognition. The performance of the proposed method is compared with existing approaches and better performance is obtained with proposed method. From experimental results it is observed that 75.32% and 84.5% recognition rate is obtained for Berlin emotional database and LDC emotional speech database respectively.

An Enhanced Latent Semantic Analysis Approach for Arabic Document Summarization

The fast-growing amount of information on the Internet makes the research in automatic document summarization very urgent. It is an effective solution for information overload. Many approaches have been proposed based on different strategies, such as latent semantic analysis (LSA). However, LSA, when applied to document summarization, has some limitations which diminish its performance. In this work, we try to overcome these limitations by applying statistic and linear algebraic approaches combined with syntactic and semantic processing of text. First, the part of speech tagger is utilized to reduce the dimension of LSA. Then, the weight of the term in four adjacent sentences is added to the weighting schemes while calculating the input matrix to take into account the word order and the syntactic relations. In addition, a new LSA-based sentence selection algorithm is proposed, in which the term description is combined with sentence description for each topic which in turn makes the generated summary more informative and diverse. To ensure the effectiveness of the proposed LSA-based sentence selection algorithm, extensive experiment on Arabic and English are done. Four datasets are used to evaluate the new model, Linguistic Data Consortium (LDC) Arabic Newswire-a corpus, Essex Arabic Summaries Corpus (EASC), DUC2002, and Multilingual MSS 2015 dataset. Experimental results on the four datasets show the effectiveness of the proposed model on Arabic and English datasets. It performs comprehensively better compared to the state-of-the-art methods.

Evaluation of an Arabic Speech Corpus of Emotions: A Perceptual and Statistical Analysis

The processing of emotion has a wide range of applications in many different fields and has become the subject of increasing interest and attention for many speech and language researchers. Speech emotion recognition systems face many challenges. One of these is the degree of naturalness of the emotions in speech corpora. To prove the ability of speakers to accurately emulate emotions and to check whether listeners could identify the intended emotion, a human perception test was designed for a new emotional speech corpus. This paper presents an exhaustive statistical and perceptual investigation of the emotional speech corpus (KSUEmotions) for Arabic King Saud University approved by the Linguistic Data Consortium. The KSUEmotions corpus was built in two phases and involved 23 native speakers (10 males and 13 females) to emulate the following five emotions: neutral, sadness, happiness, surprise, and anger. Nine listeners were participated in a blind and randomly structured human perceptual test to assess the validity of the intended emotions. Statistical tests were used to analyze the effects of speaker gender, reviewer (listener) gender, emotion type, sentence length, and the interaction between these factors. Conducted statistical tests included the two-way analysis of variance, normality, chi-square, Bonferroni, Tukey, and Mann–Whitney U tests. One of the outcomes of the study is that the speaker gender, emotion type, and interaction between emotion type and speaker gender yield significant effects on the emotion perception in this corpus.

An Event Relationship Model for Knowledge Organization and Visualization

An event is a specific occurrence involving participants, which is a typed, n-ary association of entities or other events, each identified as a participant in a specific semantic role in the event (Pyysalo et al. 2012; Linguistic Data Consortium 2005). Event types may vary across domains. Representing relationships between events can facilitate the understanding of knowledge in complex systems (such as economic systems, human body, social systems). In the simplest form, an event can be represented as Entity A -Relation- Entity B. This paper evaluates several knowledge organization and visualization models and tools, such as concept maps (Cmap), topic maps (Ontopia), network analysis models (Gephi), and ontology (Protégé), then proposes an event relationship model th at aims to integrate the strengths of these models, and can represent complex knowledge expressed in events and their relationships.

Multi-lingual geoparsing based on machine translation

Our method for multi-lingual geoparsing uses monolingual tools and resources along with machine translation and alignment to return location words in many languages. Not only does our method save the time and cost of developing geoparsers for each language separately, but also it allows the possibility of a wide range of having a wide range of language capabilities within a single interface. We evaluated our method in our LanguageBridge prototype on location named entities using newswire, broadcast news and telephone conversations in English, Arabic and Chinese data from the Linguistic Data Consortium (LDC). Our results for geoparsing Chinese and Arabic text using our multi-lingual geoparsing method are comparable to our results for geoparsing English text with our English tools. Furthermore, our experiments using our tools on machine translation approach in accuracy results on results from the same data that was translated manually, further showing the robustness of locations to machine translation.

RST Signalling Corpus: a corpus of signals of coherence relations

We present the RST Signalling Corpus (Das et al. in RST signalling corpus, LDC2015T10. https://catalog.ldc.upenn.edu/LDC2015T10 , 2015), a corpus annotated for signals of coherence relations. The corpus is developed over the RST Discourse Treebank (Carlson et al. in RST Discourse Treebank, LDC2002T07. https://catalog.ldc.upenn.edu/LDC2002T07 , 2002) which is annotated for coherence relations. In the RST Signalling Corpus, these relations are further annotated with signalling information. The corpus includes annotation not only for discourse markers which are considered to be the most typical (or sometimes the only type of) signals in discourse, but also for a wide array of other signals such as reference, lexical, semantic, syntactic, graphical and genre features as potential indicators of coherence relations. We describe the research underlying the development of the corpus and the annotation process, and provide details of the corpus. We also present the results of an inter-annotator agreement study, illustrating the validity and reproducibility of the annotation. The corpus is available through the Linguistic Data Consortium, and can be used to investigate the psycholinguistic mechanisms behind the interpretation of relations through signalling, and also to develop discourse-specific computational systems such as discourse parsing applications.

Facial Expression Recognition with Faster R-CNN

In order to avoid the complex explicit feature extraction process and the problem of low-level data operation involved in traditional facial expression recognition, we proposed a method of Faster R-CNN (Faster Regions with Convolutional Neural Network Features) for facial expression recognition in this paper. Firstly, the facial expression image is normalized and the implicit features are extracted by using the trainable convolution kernel. Then, the maximum pooling is used to reduce the dimensions of the extracted implicit features. After that, RPNs (Region Proposal Networks) is used to generate high-quality region proposals, which are used by Faster R-CNN for detection. Finally, the Softmax classifier and regression layer is used to classify the facial expressions and predict boundary box of the test sample, respectively. The dataset is provided by Chinese Linguistic Data Consortium (CLDC), which is composed of multimodal emotional audio and video data. Experimental results show the performance and the generalization ability of the Faster R-CNN for facial expression recognition. The value of the mAP is around 0.82.

Multistage data selection-based unsupervised speaker adaptation for personalized speech emotion recognition

This paper proposes an efficient speech emotion recognition (SER) approach that utilizes personal voice data accumulated on personal devices. A representative weakness of conventional SER systems is the user-dependent performance induced by the speaker independent (SI) acoustic model framework. But, handheld communications devices such as smartphones provide a collection of individual voice data, thus providing suitable conditions for personalized SER that is more enhanced than the SI model framework. By taking advantage of personal devices, we propose an efficient personalized SER scheme employing maximum likelihood linear regression (MLLR), a representative speaker adaptation technique. To further advance the conventional MLLR technique for SER tasks, the proposed approach selects useful data that convey emotionally discriminative acoustic characteristics and uses only those data for adaptation. For reliable data selection, we conduct multistage selection using a log-likelihood distance-based measure and a universal background model. On SER experiments based on a Linguistic Data Consortium emotional speech corpus, our approach exhibited superior performance when compared to conventional adaptation techniques as well as the SI model framework.