High-Order Concept Associations Mining and Inferential Language Modeling for Online Review Spam Detection

Despite many incidents about fake online consumer reviews have been reported, very few studies have been conducted to date to examine the trustworthiness of online consumer reviews. One of the reasons is the lack of an effective computational method to separate the untruthful reviews (i.e., spam) from the legitimate ones (i.e., ham) given the fact that prominent spam features are often missing in online reviews. The main contribution of our research work is the development of a novel review spam detection method which is underpinned by an unsupervised inferential language modeling framework. Another contribution of this work is the development of a high-order concept association mining method which provides the essential term association knowledge to bootstrap the performance for untruthful review detection. Our experimental results confirm that the proposed inferential language model equipped with high-order concept association knowledge is effective in untruthful review detection when compared with other baseline methods.

Text mining and probabilistic language modeling for online review spam detecting

In the era of Web 2.0, huge volumes of consumer reviews are posted to the Internet every day. Manual approaches to detecting and analyzing fake reviews (i.e., spam) are not practical due to the problem of information overload. However, the design and development of automated methods of detecting fake reviews is a challenging research problem. The main reason is that fake reviews are specifically composed to mislead readers, so they may appear the same as legitimate reviews (i.e., ham). As a result, discriminatory features that would enable individual reviews to be classified as spam or ham may not be available. Guided by the design science research methodology, the main contribution of this study is the design and instantiation of novel computational models for detecting fake reviews. In particular, a novel text mining model is developed and integrated into a semantic language model for the detection of untruthful reviews. The models are then evaluated based on a real-world dataset collected from amazon.com. The results of our experiments confirm that the proposed models outperform other well-known baseline models in detecting fake reviews. To the best of our knowledge, the work discussed in this article represents the first successful attempt to apply text mining methods and semantic language models to the detection of fake consumer reviews. A managerial implication of our research is that firms can apply our design artifacts to monitor online consumer reviews to develop effective marketing or product design strategies based on genuine consumer feedback posted to the Internet.

Facilitating query decomposition in query language modeling by association rule mining using multiple sliding windows

This paper presents a novel framework to further advance the recent trend of using query decomposition and high-order term relationships in query language modeling, which takes into account terms implicitly associated with different subsets of query terms. Existing approaches, most remarkably the language model based on the Information Flow method are however unable to capture multiple levels of associations and also suffer from a high computational overhead. In this paper, we propose to compute association rules from pseudo feedback documents that are segmented into variable length chunks via multiple sliding windows of different sizes. Extensive experiments have been conducted on various TREC collections and our approach significantly outperforms a baseline Query Likelihood language model, the Relevance Model and the Information Flow model.

Suite of tools for statistical N-gram language modeling for pattern mining in whole genome sequences

Genome sequences contain a number of patterns that have biomedical significance. Repetitive sequences of various kinds are a primary component of most of the genomic sequence patterns. We extended the suffix-array based Biological Language Modeling Toolkit to compute n-gram frequencies as well as n-gram language-model based perplexity in windows over the whole genome sequence to find biologically relevant patterns. We present the suite of tools and their application for analysis on whole human genome sequence.

An effective approach to verbose queries using a limited dependencies language model

Intuitively, any `bag of words' approach in IR should benefit from taking term dependencies into account. Unfortunately, for years the results of exploiting such dependencies have been mixed or inconclusive. To improve the situation, this paper shows how the natural language properties of the target documents can be used to transform and enrich the term dependencies to more useful statistics. This is done in three steps. The term co-occurrence statistics of queries and documents are each represented by a Markov chain. The paper proves that such a chain is ergodic, and therefore its asymptotic behavior is unique, stationary, and independent of the initial state. Next, the stationary distribution is taken to model queries and documents, rather than their initial distri- butions. Finally, ranking is achieved following the customary language modeling paradigm. The main contribution of this paper is to argue why the asymptotic behavior of the document model is a better representation then just the document's initial distribution. A secondary contribution is to investigate the practical application of this representation in case the queries become increasingly verbose. In the experiments (based on Lemur's search engine substrate) the default query model was replaced by the stable distribution of the query. Just modeling the query this way already resulted in significant improvements over a standard language model baseline. The results were on a par or better than more sophisticated algorithms that use fine-tuned parameters or extensive training. Moreover, the more verbose the query, the more effective the approach seems to become.

Query expansion using term relationships in language models for information retrieval

Language Modeling (LM) has been successfully applied to Information Retrieval (IR). However, most of the existing LM approaches only rely on term occurrences in documents, queries and document collections. In traditional unigram based models, terms (or words) are usually considered to be independent. In some recent studies, dependence models have been proposed to incorporate term relationships into LM, so that links can be created between words in the same sentence, and term relationships (e.g. synonymy) can be used to expand the document model. In this study, we further extend this family of dependence models in the following two ways: (1) Term relationships are used to expand query model instead of document model, so that query expansion process can be naturally implemented; (2) We exploit more sophisticated inferential relationships extracted with Information Flow (IF). Information flow relationships are not simply pairwise term relationships as those used in previous studies, but are between a set of terms and another term. They allow for context-dependent query expansion. Our experiments conducted on TREC collections show that we can obtain large and significant improvements with our approach. This study shows that LM is an appropriate framework to implement effective query expansion.

An effective approach to verbose queries using a limited dependencies language model

Intuitively, any ‘bag of words’ approach in IR should benefit from taking term dependencies into account. Unfortunately, for years the results of exploiting such dependencies have been mixed or inconclusive. To improve the situation, this paper shows how the natural language properties of the target documents can be used to transform and enrich the term dependencies to more useful statistics. This is done in three steps. The term co-occurrence statistics of queries and documents are each represented by a Markov chain. The paper proves that such a chain is ergodic, and therefore its asymptotic behavior is unique, stationary, and independent of the initial state. Next, the stationary distribution is taken to model queries and documents, rather than their initial distributions. Finally, ranking is achieved following the customary language modeling paradigm. The main contribution of this paper is to argue why the asymptotic behavior of the document model is a better representation then just the document’s initial distribution. A secondary contribution is to investigate the practical application of this representation in case the queries become increasingly verbose. In the experiments (based on Lemur’s search engine substrate) the default query model was replaced by the stable distribution of the query. Just modeling the query this way already resulted in significant improvements over a standard language model baseline. The results were on a par or better than more sophisticated algorithms that use fine-tuned parameters or extensive training. Moreover, the more verbose the query, the more effective the approach seems to become.

An aspect query language model based on query decomposition and high-order contextual term associations

In information retrieval (IR) research, more and more focus has been placed on optimizing a query language model by detecting and estimating the dependencies between the query and the observed terms occurring in the selected relevance feedback documents. In this paper, we propose a novel Aspect Language Modeling framework featuring term association acquisition, document segmentation, query decomposition, and an Aspect Model (AM) for parameter optimization. Through the proposed framework, we advance the theory and practice of applying high-order and context-sensitive term relationships to IR. We first decompose a query into subsets of query terms. Then we segment the relevance feedback documents into chunks using multiple sliding windows. Finally we discover the higher order term associations, that is, the terms in these chunks with high degree of association to the subsets of the query. In this process, we adopt an approach by combining the AM with the Association Rule (AR) mining. In our approach, the AM not only considers the subsets of a query as “hidden” states and estimates their prior distributions, but also evaluates the dependencies between the subsets of a query and the observed terms extracted from the chunks of feedback documents. The AR provides a reasonable initial estimation of the high-order term associations by discovering the associated rules from the document chunks. Experimental results on various TREC collections verify the effectiveness of our approach, which significantly outperforms a baseline language model and two state-of-the-art query language models namely the Relevance Model and the Information Flow model

Pruned Universal Symbol Sequences for LZW based Language identification

We present a improved language modeling technique for Lempel-Ziv-Welch (LZW) based LID scheme. The previous approach to LID using LZW algorithm prepares the language pattern table using LZW algorithm. Because of the sequential nature of the LZW algorithm, several language specific patterns of the language were missing in the pattern table. To overcome this, we build a universal pattern table, which contains all patterns of different length. For each language it's corresponding language specific pattern table is constructed by retaining the patterns of the universal table whose frequency of appearance in the training data is above the threshold.This approach reduces the classification score (Compression Ratio [LZW-CR] or the weighted discriminant score[LZW-WDS]) for non native languages and increases the LID performance considerably.

LZW Based Distance Measures for Spoken Language Identification

We present a new approach to spoken language modeling for language identification (LID) using the Lempel-Ziv-Welch (LZW) algorithm. The LZW technique is applicable to any kind of tokenization of the speech signal. Because of the efficiency of LZW algorithm to obtain variable length symbol strings in the training data, the LZW codebook captures the essentials of a language effectively. We develop two new deterministic measures for LID based on the LZW algorithm namely: (i) Compression ratio score (LZW-CR) and (ii) weighted discriminant score (LZW-WDS). To assess these measures, we consider error-free tokenization of speech as well as artificially induced noise in the tokenization. It is shown that for a 6 language LID task of OGI-TS database with clean tokenization, the new model (LZW-WDS) performs slightly better than the conventional bigram model. For noisy tokenization, which is the more realistic case, LZW-WDS significantly outperforms the bigram technique

Discriminative language model adaptation for Mandarin broadcast speech transcription and translation

This paper investigates unsupervised test-time adaptation of language models (LM) using discriminative methods for a Mandarin broadcast speech transcription and translation task. A standard approach to adapt interpolated language models to is to optimize the component weights by minimizing the perplexity on supervision data. This is a widely made approximation for language modeling in automatic speech recognition (ASR) systems. For speech translation tasks, it is unclear whether a strong correlation still exists between perplexity and various forms of error cost functions in recognition and translation stages. The proposed minimum Bayes risk (MBR) based approach provides a flexible framework for unsupervised LM adaptation. It generalizes to a variety of forms of recognition and translation error metrics. LM adaptation is performed at the audio document level using either the character error rate (CER), or translation edit rate (TER) as the cost function. An efficient parameter estimation scheme using the extended Baum-Welch (EBW) algorithm is proposed. Experimental results on a state-of-the-art speech recognition and translation system are presented. The MBR adapted language models gave the best recognition and translation performance and reduced the TER score by up to 0.54% absolute. © 2007 IEEE.

Extracting and exploiting word relationships for information retrieval

Thèse numérisée par la Division de la gestion de documents et des archives de l'Université de Montréal

Relating Dependent Terms in Information Retrieval

Les moteurs de recherche font partie de notre vie quotidienne. Actuellement, plus d’un tiers de la population mondiale utilise l’Internet. Les moteurs de recherche leur permettent de trouver rapidement les informations ou les produits qu'ils veulent. La recherche d'information (IR) est le fondement de moteurs de recherche modernes. Les approches traditionnelles de recherche d'information supposent que les termes d'indexation sont indépendants. Pourtant, les termes qui apparaissent dans le même contexte sont souvent dépendants. L’absence de la prise en compte de ces dépendances est une des causes de l’introduction de bruit dans le résultat (résultat non pertinents). Certaines études ont proposé d’intégrer certains types de dépendance, tels que la proximité, la cooccurrence, la contiguïté et de la dépendance grammaticale. Dans la plupart des cas, les modèles de dépendance sont construits séparément et ensuite combinés avec le modèle traditionnel de mots avec une importance constante. Par conséquent, ils ne peuvent pas capturer correctement la dépendance variable et la force de dépendance. Par exemple, la dépendance entre les mots adjacents "Black Friday" est plus importante que celle entre les mots "road constructions". Dans cette thèse, nous étudions différentes approches pour capturer les relations des termes et de leurs forces de dépendance. Nous avons proposé des méthodes suivantes: ─ Nous réexaminons l'approche de combinaison en utilisant différentes unités d'indexation pour la RI monolingue en chinois et la RI translinguistique entre anglais et chinois. En plus d’utiliser des mots, nous étudions la possibilité d'utiliser bi-gramme et uni-gramme comme unité de traduction pour le chinois. Plusieurs modèles de traduction sont construits pour traduire des mots anglais en uni-grammes, bi-grammes et mots chinois avec un corpus parallèle. Une requête en anglais est ensuite traduite de plusieurs façons, et un score classement est produit avec chaque traduction. Le score final de classement combine tous ces types de traduction. Nous considérons la dépendance entre les termes en utilisant la théorie d’évidence de Dempster-Shafer. Une occurrence d'un fragment de texte (de plusieurs mots) dans un document est considérée comme représentant l'ensemble de tous les termes constituants. La probabilité est assignée à un tel ensemble de termes plutôt qu’a chaque terme individuel. Au moment d’évaluation de requête, cette probabilité est redistribuée aux termes de la requête si ces derniers sont différents. Cette approche nous permet d'intégrer les relations de dépendance entre les termes. Nous proposons un modèle discriminant pour intégrer les différentes types de dépendance selon leur force et leur utilité pour la RI. Notamment, nous considérons la dépendance de contiguïté et de cooccurrence à de différentes distances, c’est-à-dire les bi-grammes et les paires de termes dans une fenêtre de 2, 4, 8 et 16 mots. Le poids d’un bi-gramme ou d’une paire de termes dépendants est déterminé selon un ensemble des caractères, en utilisant la régression SVM. Toutes les méthodes proposées sont évaluées sur plusieurs collections en anglais et/ou chinois, et les résultats expérimentaux montrent que ces méthodes produisent des améliorations substantielles sur l'état de l'art.

Dificultades de comprensión lingüística oral, propuesta de mejora utilizando las nuevas tecnologías : diálogo hombre-máquina mediante instrucciones expresadas en lenguaje natural.

Responder cómo se procesa el lenguaje, cómo funcionan todos los elementos que intervienen en la comprensión y en qué orden se produce el procesamiento lingüístico. Alumnos de ESO, que no presentan discapacidad auditiva. El grupo experimental lo compone 31 chicos y 12 niñas que presentan dificultades en Lengua, algunos de ellos también tienen problemas de aprendizaje en Matemáticas y Lengua inglesa. Se realizan dos pruebas. La primera trata de comprensión oral. Reciben un cuadernillo cada uno. Disponen de 25 minutos. Los datos personales es lo último que deben escribir. Si no escuchan bien, lo indican en el cuadernillo y así se controla la falta de comprensión por deficiencias de sonido. Se les pone una grabación tres veces. Durante las grabaciones se controlan las diferencias acústicas entre los que están en la primera fila y la última. Los alumnos contestan a las preguntas. A los que presentan problemas con las definiciones se les pide que rellenen la última hoja para comprobar si conocen el significado, no su capacidad de expresión. El segundo cuadernillo lo reciben al acabar todo el grupo. Disponen de tiempo ilimitado. Si no conocen una palabra se les explica el significado. Finalmente se les pasa una prueba de memoria auditiva inmediata . Se pretende controlar la variable 'memoria' y estudiar su incidencia en la prueba. La segunda prueba consiste en originar un modelo de lenguaje utilizando el mismo texto presentado a los estudiantes. También se pretende conocer lo que pasa si se introducen oraciones incompletas para rellenar por los alumnos. La única información que dispone el ordenador es la señal vocal y con ella realiza el modelo de lenguaje. Grabadora mono portátil, cinta de casete, cuadernillo de respuesta de comprensión oral, cuadernillo de respuesta de estrategias de comprensión utilizada, cuaderno de respuestas de estrategias de procedimiento de comprensión, hoja de respuestas para la prueba de memoria, programa SPSS y Excel para análisis de datos. Para la segunda prueba los materiales son: la grabadora mono portátil Panasonic, cinta casete, reconocedor Via Voice 98, Pentium III, tarjeta de sonido, C.M.U. Statistical Language Modeling Tool Kit, Programa tex2wfreq, text2idngram, idngram21m,evallm. Para la primera prueba se confecciona un diseño experimental multivariado; las variables fueron: memoria, comprensión auditiva y estrategias utilizadas para comprender. Las variables contaminadoras: experimentador, material, condiciones acústicas, centro educativo, nivel socioeconómico y edad. Éstas se controlan por igualación. Las variables organísmicas y el sexo se controlan aleatoriamente. La memoria auditiva tuvo que ser controlada a través de un análisis de covarianza. En la segunda prueba, la variable fue la comprensión lingüística oral, para después establecer una comparación. Los resultados de la primera prueba revelan que las correlaciones que se obtienen entre las variables analizadas son independientes y arrojan diferencias entre el grupo experimental y el de control. Se encuentran puntuaciones más altas en los sujetos sin dificultades en memoria y comprensión. No hay diferencias entre los dos grupos en estrategias de comprensión. Los resultados obtenidos en la fase de evaluación de la segunda prueba indican que ninguna respuesta resulta elegida correctamente por lo que no se puede realizar ninguna comparación . Parece que la muestra utiliza el mismo modelo para comprender, todos utilizan las mismas estrategias, las diferencias son cuantitativas y debidas a variables organísmicas, entre ellas, la memoria. La falta de vocabulario es la primera dificultad en el grupo con dificultades, la falta de memoria impide corregir palabras mal pronunciadas, buscar conocimientos previos y relacionar ideas en su memoria a largo plazo. Son también incapaces de encontrar la idea principal. La comprensión es tan lenta que no pueden procesar. Se demuestra que los programas informáticos imitan al hombre a niveles elementales, en Tecnología del Habla se utilizan prioritariamente modelos semánticos.