Search results for: unstructured text

Authors: Awatef Boubakri, Khaled Jebahi

Abstract:

Using Text World Theory (TWT), we attempted to understand how mental representations (text worlds) and perceptions can be construed by readers of Quranic texts. To this end, Surah Al-Kahf was purposefully selected given the fact that while each of its stories is narrated, different levels of discourse intervene, which might result in a confused reader who might find it hard to keep track of which discourse he or she is processing. This surah was studied using specifically-designed text-world diagrams. The findings suggest that TWT can be used to help solve problems of ambiguity at the level of discourse in Quranic texts and to help construct a thinking reader whose cognitive constructs (text worlds / mental representations) are built through reflecting on the various and often changing components of discourse world, text world, and sub-worlds.

Keywords: Al-Kahf, Surah, cognitive, processing, discourse

Procedia PDF Downloads 53

Authors: B. S. Harish, M. B. Revanasiddappa

Abstract:

Due to rapid growth of text documents in digital form, automated text classification has become an important research in the last two decades. The major challenge of text document representations are high dimension, sparsity, volume and semantics. Since the terms are only features that can be found in documents, selection of good terms (features) plays an very important role. In text classification, feature selection is a strategy that can be used to improve classification effectiveness, computational efficiency and accuracy. In this paper, we present a quantitative analysis of most widely used feature selection (FS) methods, viz. Term Frequency-Inverse Document Frequency (tf
idf ), Mutual Information (MI), Information Gain (IG), CHISquare (x2), Term Frequency-Relevance Frequency (tf
rf ), Term Strength (TS), Ambiguity Measure (AM) and Symbolic Feature Selection (SFS) to classify text documents. We evaluated all the feature selection methods on standard datasets like 20 Newsgroups, 4 University dataset and Reuters-21578.

Keywords: classifiers, feature selection, text classification

Procedia PDF Downloads 423

Authors: Yu Zhang, Pedro Desouza

Abstract:

Since big data has become substantially more accessible and manageable due to the development of powerful tools for dealing with unstructured data, people are eager to mine information from social media resources that could not be handled in the past. Sentiment analysis, as a novel branch of text mining, has in the last decade become increasingly important in marketing analysis, customer risk prediction and other fields. Scientists and researchers have undertaken significant work in creating and improving their sentiment models. In this paper, we present a concept of selecting appropriate classifiers based on the features and qualities of data sources by comparing the performances of five classifiers with three popular social media data sources: Twitter, Amazon Customer Reviews, and Movie Reviews. We introduced a couple of innovative models that outperform traditional sentiment classifiers for these data sources, and provide insights on how to further improve the predictive power of sentiment analysis. The modelling and testing work was done in R and Greenplum in-database analytic tools.

Keywords: sentiment analysis, social media, Twitter, Amazon, data mining, machine learning, text mining

Procedia PDF Downloads 325

Authors: Jaehyung An, Sungjoo Lee

Abstract:

Patent documents constitute an up-to-date and reliable source of knowledge for reflecting technological advance, so patent analysis has been widely used for identification of technological trends and formulation of technology strategies. But, identifying technological information from patent data entails some limitations such as, high cost, complexity, and inconsistency because it rely on the expert’ knowledge. To overcome these limitations, researchers have applied to a quantitative analysis based on the keyword technique. By using this method, you can include a technological implication, particularly patent documents, or extract a keyword that indicates the important contents. However, it only uses the simple-counting method by keyword frequency, so it cannot take into account the sematic relationship with the keywords and sematic information such as, how the technologies are used in their technology area and how the technologies affect the other technologies. To automatically analyze unstructured technological information in patents to extract the semantic information, it should be transformed into an abstracted form that includes the technological key concepts. Specific sentence structure ‘SAO’ (subject, action, object) is newly emerged by representing ‘key concepts’ and can be extracted by NLP (Natural language processor). An SAO structure can be organized in a problem-solution format if the action-object (AO) states that the problem and subject (S) form the solution. In this paper, we propose the new methodology that can extract the SAO structure through technical elements extracting rules. Although sentence structures in the patents text have a unique format, prior studies have depended on general NLP (Natural language processor) applied to the common documents such as newspaper, research paper, and twitter mentions, so it cannot take into account the specific sentence structure types of the patent documents. To overcome this limitation, we identified a unique form of the patent sentences and defined the SAO structures in the patents text data. There are four types of technical elements that consist of technology adoption purpose, application area, tool for technology, and technical components. These four types of sentence structures from patents have their own specific word structure by location or sequence of the part of speech at each sentence. Finally, we developed algorithms for extracting SAOs and this result offer insight for the technology innovation process by providing different perspectives of technology.

Keywords: NLP, patent analysis, SAO, semantic-analysis

Procedia PDF Downloads 241

Authors: Sharun Akter Khushbu

Abstract:

‘SOS’ addresses text summary such as feasibility study and allows more comprehensive methods on text of language resources. Resources language has been exploited by the importance of text documental procedure. Throughout this key idea will come out a machine interpreter called an SOS that has built an argumentative as an employed model is LSTM-CNN(long short-term memory- recurrent neural network). Summarization of Bengali text formulated by the information of latent structure instead of brief input string counting as text. Text summarization is the proper utilization of optimal solutions being time reduction, and easy interpretation whenever human-generated summary and machine targeted summary remain similar and without degrading the semantic summarization quality. According to the problem affirmation key idea has advanced an algorithm with the method of encoder and decoder describing a sequential structure that is rigorously connected with actual predicted and meaningful output. Regarding the seq2seq approach aimed in the future with high semantic summarization similarity on behalf of the large data samples that are also enlisted by the method. Thus, the SOS method assigns a discriminator over Bengali text documents where encoded input sequences such as summary and decoded the targeted summary of gist will be an error-free machine.

Keywords: LSTM-CNN, NN, SOS, text summarization

Procedia PDF Downloads 41

Authors: Shen Jian, Hu Jie, Qi Jin, Liu Wei Jie, Chen Ji Yi, Peng Ying Hong

Abstract:

In order to settle the problem of acquiring case in biological related to design problems, a biometrics instance acquisition method based on text mining is presented. Through the construction of corpus text vector space and knowledge mining, the feature selection, similarity measure and case retrieval method of text in the field of biology are studied. First, we establish a vector space model of the corpus in the biological field and complete the preprocessing steps. Then, the corpus is retrieved by using the vector space model combined with the functional keywords to obtain the biological domain examples related to the design problems. Finally, we verify the validity of this method by taking the example of text.

Keywords: text mining, vector space model, feature selection, biologically inspired design

Procedia PDF Downloads 227

Authors: Bashar Albaalbaki, Roger E. Khayat

Abstract:

This work is a comparative study on the effect of Delaunay triangulation algorithms on discretization error for conduction-convection conservation problems. A structured triangulation and many unstructured Delaunay triangulations using three popular algorithms for node placement strategies are used. The numerical method employed is the vertex-centered finite volume method. It is found that when the computational domain can be meshed using a structured triangulation, the discretization error is lower for structured triangulations compared to unstructured ones for only low Peclet number values, i.e. when conduction is dominant. However, as the Peclet number is increased and convection becomes more significant, the unstructured triangulations reduce the discretization error. Also, no statistical correlation between triangulation angle extremums and the discretization error is found using 200 samples of randomly generated Delaunay and non-Delaunay triangulations. Thus, the angle extremums cannot be an indicator of the discretization error on their own and need to be combined with other triangulation quality measures, which is the subject of further studies.

Keywords: conduction-convection problems, Delaunay triangulation, discretization error, finite volume method

Procedia PDF Downloads 63

Authors: Omid Shahmirzadi, Adam Lugowski, Kenneth Younge

Abstract:

Automatic measurement of semantic text similarity is an important task in natural language processing. In this paper, we evaluate the performance of different vector space models to perform this task. We address the real-world problem of modeling patent-to-patent similarity and compare TFIDF (and related extensions), topic models (e.g., latent semantic indexing), and neural models (e.g., paragraph vectors). Contrary to expectations, the added computational cost of text embedding methods is justified only when: 1) the target text is condensed; and 2) the similarity comparison is trivial. Otherwise, TFIDF performs surprisingly well in other cases: in particular for longer and more technical texts or for making finer-grained distinctions between nearest neighbors. Unexpectedly, extensions to the TFIDF method, such as adding noun phrases or calculating term weights incrementally, were not helpful in our context.

Keywords: big data, patent, text embedding, text similarity, vector space model

Procedia PDF Downloads 140

Authors: Aliya Grig, Konstantin Sokolov, Igor Shatalin

Abstract:

The report describes the methodology for collecting data and building an ML model for determining the personality psychotype using profiling and personality traits methods based on several short messages of a user communicating on an arbitrary topic with a chitchat bot. In the course of the experiments, the minimum amount of text was revealed to confidently determine aspects of personality. Model accuracy - 85%. Users' language of communication is English. AI for a personalized communication with a user based on his mood, personality, and current emotional state. Features investigated during the research: personalized communication; providing empathy; adaptation to a user; predictive analytics. In the report, we describe the processes that captures both structured and unstructured data pertaining to a user in large quantities and diverse forms. This data is then effectively processed through ML tools to construct a knowledge graph and draw inferences regarding users of text messages in a comprehensive manner. Specifically, the system analyzes users' behavioral patterns and predicts future scenarios based on this analysis. As a result of the experiments, we provide for further research on training AI models to be empathetic, creating personalized communication for a user

Keywords: AI, empathetic, chatbot, AI models

Procedia PDF Downloads 60

Authors: Roman Graf, Sergiu Gordea, Heather M. Ryan

Abstract:

This paper presents an approach for the classification of an unstructured format description for identification of file formats. The main contribution of this work is the employment of data mining techniques to support file format selection with just the unstructured text description that comprises the most important format features for a particular organisation. Subsequently, the file format indentification method employs file format classifier and associated configurations to support digital preservation experts with an estimation of required file format. Our goal is to make use of a format specification knowledge base aggregated from a different Web sources in order to select file format for a particular institution. Using the naive Bayes method, the decision support system recommends to an expert, the file format for his institution. The proposed methods facilitate the selection of file format and the quality of a digital preservation process. The presented approach is meant to facilitate decision making for the preservation of digital content in libraries and archives using domain expert knowledge and specifications of file formats. To facilitate decision-making, the aggregated information about the file formats is presented as a file format vocabulary that comprises most common terms that are characteristic for all researched formats. The goal is to suggest a particular file format based on this vocabulary for analysis by an expert. The sample file format calculation and the calculation results including probabilities are presented in the evaluation section.

Keywords: data mining, digital libraries, digital preservation, file format

Procedia PDF Downloads 472

Authors: Mahsa Khani Oushani, Mohammad Kazem Hasanvand

Abstract:

Text and image have always been two important components in Iranian layout. The interactive connection between text and image has shaped the art of book design with multiple patterns. In this research, first the structure and visual elements in the research data were analyzed and then the position of the text element and the image element in relation to each other based on Roland Barthes theory on the three theories of text and image, were studied and analyzed and the results were compared, and interpreted. The purpose of this study is to investigate the pattern of text and image in the works of Kamaluddin Behzad based on three Roland Barthes communication theories, 1. Descriptive communication, 2. Reference communication, 3. Matched communication. The questions of this research are what is the relationship between text and image in Behzad's works? And how is it defined according to Roland Barthes theory? The method of this research has been done with a structuralist approach with a descriptive-analytical method in a library collection method. The information has been collected in the form of documents (library) and is a tool for collecting online databases. Findings show that the dominant element in Behzad's drawings is with the image and has created a reference relationship in the layout of the drawings, but in some cases it achieves a different relationship that despite the preference of the image on the page, the text is dispersed proportionally on the page and plays a more active role, played within the image. The text and the image support each other equally on the page; Roland Barthes equates this connection.

Keywords: text, image, Kamaluddin Behzad, Roland Barthes, communication theory

Procedia PDF Downloads 160

Authors: Dong Seop Lee, Byung Sik Kim

Abstract:

In the 4th Industrial Revolution, various intelligent technologies have been developed in many fields. These artificial intelligence technologies are applied in various services, including disaster management. Disaster information management does not just support disaster work, but it is also the foundation of smart disaster management. Furthermore, it gets historical disaster information using artificial intelligence technology. Disaster information is one of important elements of entire disaster cycle. Disaster information management refers to the act of managing and processing electronic data about disaster cycle from its’ occurrence to progress, response, and plan. However, information about status control, response, recovery from natural and social disaster events, etc. is mainly managed in the structured and unstructured form of reports. Those exist as handouts or hard-copies of reports. Such unstructured form of data is often lost or destroyed due to inefficient management. It is necessary to manage unstructured data for disaster information. In this paper, the Optical Character Recognition approach is used to convert handout, hard-copies, images or reports, which is printed or generated by scanners, etc. into electronic documents. Following that, the converted disaster data is organized into the disaster code system as disaster information. Those data are stored in the disaster database system. Gathering and creating disaster information based on Optical Character Recognition for unstructured data is important element as realm of the smart disaster management. In this paper, Korean characters were improved to over 90% character recognition rate by using upgraded OCR. In the case of character recognition, the recognition rate depends on the fonts, size, and special symbols of character. We improved it through the machine learning algorithm. These converted structured data is managed in a standardized disaster information form connected with the disaster code system. The disaster code system is covered that the structured information is stored and retrieve on entire disaster cycle such as historical disaster progress, damages, response, and recovery. The expected effect of this research will be able to apply it to smart disaster management and decision making by combining artificial intelligence technologies and historical big data.

Keywords: disaster information management, unstructured data, optical character recognition, machine learning

Procedia PDF Downloads 95

Authors: Jaspreet Singh, Gurvinder Singh, Prabhsimran Singh, Rajinder Singh, Prithvipal Singh, Karanjeet Singh Kahlon, Ravinder Singh Sawhney

Abstract:

Morphological evaluation of Indian languages is one of the burgeoning fields in the area of Natural Language Processing (NLP). The evaluation of a language is an eminent task in the era of information retrieval and text mining. The extraction and classification of knowledge from text can be exploited for sentiment analysis and morphological evaluation. This study coalesce morphological evaluation and sentiment analysis for the task of classification of farmer suicide cases reported in Punjab state of India. The pre-processing of Punjabi text involves morphological evaluation and normalization of Punjabi word tokens followed by the training of proposed model using deep learning classification on Punjabi language text extracted from online Punjabi news reports. The class-wise accuracies of sentiment prediction for four negatively oriented classes of farmer suicide cases are 93.85%, 88.53%, 83.3%, and 95.45% respectively. The overall accuracy of sentiment classification obtained using proposed framework on 275 Punjabi text documents is found to be 90.29%.

Keywords: deep neural network, farmer suicides, morphological processing, punjabi text, sentiment analysis

Procedia PDF Downloads 286

Authors: Sandeep Sharma, Sarabjit Singh

Abstract:

As per the user demand and growth trends of large free data the storage solutions are now becoming more challenge-able to protect, store and to retrieve data. The days are not so far when the storage companies and organizations are start saying 'no' to store our valuable data or they will start charging a huge amount for its storage and protection. On the other hand as per the environmental conditions it becomes challenge-able to maintain and establish new data warehouses and data centers to protect global warming threats. A challenge of small data is over now, the challenges are big that how to manage the exponential growth of data. In this paper we have analyzed the growth trend of big data and its future implications. We have also focused on the impact of the unstructured data on various concerns and we have also suggested some possible remedies to streamline big data.

Keywords: big data, unstructured data, volume, variety, velocity

Procedia PDF Downloads 511

Authors: Muhammad Fairul Azreen Mohd Zahid

Abstract:

Speech, text, and movement intensify aspects of creating choreography by connecting with emotional entanglements, tradition, literature, and other texts. This research focuses on the practice as research that will prioritise the choreography process as an inquiry approach. With the driven context, the study intervenes in critical conjunctions of choreographic theory, bringing together new reflections on the moving body, spaces of action, as well as intertextuality between text and movements in making choreography. Throughout the process, the researcher will introduce the level of deliberation from speech through movements and text to express emotion within a narrative context of an “illocutionary act.” This practice as research will produce a different meaning from the “utterance text” to “utterance movements” in the perspective of speech acts theory by J.L Austin based on fragmented text from “pidato adat” which has been used as opening speech in Randai. Looking at the theory of deconstruction by Jacque Derrida also will give a different meaning from the text. Nevertheless, the process of creating the choreography will also help to lay the basic normative structure implicit in “constative” (statement text/movement) and “performative” (command text/movement). Through this process, the researcher will also look at several methods of using text from two works by Joseph Gonzales, “Becoming King-The Pakyung Revisited” and Crystal Pite's “The Statement,” as references to produce different methods in making choreography. The perspective from the semiotic foundation will support how occurrences within dance discourses as texts through a semiotic lens. The method used in this research is qualitative, which includes an interview and simulation of the concept to get an outcome.

Keywords: intertextuality, choreography, speech act, performative, deconstruction

Procedia PDF Downloads 65

Authors: Sara Zadunaisky Ehrlich, Batia Seroussi, Anat Stavans

Abstract:

Text structure is a parameter of text quality. This study investigated the structure of written argumentative texts produced by elementary school age children. We set two objectives: to identify and trace the structural components of the argumentative texts and to investigate whether reading comprehension skills were correlated with text structure. 293 school children from 2nd to 5th grades were asked to write two argumentative texts about informal or everyday life controversial topics and completed two reading tasks that targeted different levels of text comprehension. The findings indicated, on the one hand, significant developmental differences between mature and more novice writers in terms of text length and mean proportion of clauses produced for a better elaboration of the different text components. On the other hand, with certain fluctuations, no meaningful differences were found in terms of presence of text structure: at all grade levels, elementary school children produced the basic and minimal structure that included the writer's argument and reasons or arguments' supports. Counter-arguments were scarce even in the upper grades. While the children captured that essentially an argument must be justified, the more the number of supports produced, the fewer the clauses the children produced. Last, weak to mild relations were found between reading comprehension and argumentative text structure. Nevertheless, children who scored higher on sophisticated questions that require inferential or world knowledge displayed more elaborated structures in terms of text length and size of supports to the writer's argument. These findings indicate how school-age children perceive the basic template of an argument with future implications regarding how to elaborate written arguments.

Keywords: argumentative text, text structure, elementary school children, written argumentations

Procedia PDF Downloads 123

Authors: Chamindi Dilkushi Senaratne

Abstract:

Communicating via a text or an SMS (Short Message Service) has become an integral part of our daily lives. With the increase in the use of mobile phones, text messaging has become a genre by itself worth researching and studying. It is undoubtedly a major phenomenon revealing language change. This paper attempts to describe the morphological processes of text language of urban bilinguals in Sri Lanka. It will be a typological study based on 500 English text messages collected from urban bilinguals residing in Colombo. The messages are selected by categorizing the deviant forms of language use apparent in text messages. These stylistic deviations are a deliberate skilled performance by the users of the language possessing an in-depth knowledge of linguistic systems to create new words and thereby convey their linguistic identity and individual and group solidarity via the message. The findings of the study solidifies arguments that the manipulation of language in text messages is both creative and appropriate. In addition, code mixing theories will be used to identify how existing morphological processes are adapted by bilingual users in Sri Lanka when texting. The study will reveal processes such as omission, initialism, insertion and alternation in addition to other identified linguistic features in text language. The corpus reveals the most common morphological processes used by Sri Lankan urban bilinguals when sending texts.

Keywords: bilingual, deviations, morphology, texts

Procedia PDF Downloads 243

Authors: Oussama Hafferssas, Hiba Benyahia, Amina Madani, Nassima Zeriri

Abstract:

Although data nowadays has multiple forms; from text to images, and from audio to videos, yet text is still the most used one at a public level. At an academical and research level, and unlike other forms, text can be considered as the easiest form to process. Therefore, a brunch of Data Mining researches has been always under its shadow, called "Text Mining". Its concept is just like data mining’s, finding valuable patterns in data, from large collections and tremendous volumes of data, in this case: Text. Named entity recognition (NER) is one of Text Mining’s disciplines, it aims to extract and classify references such as proper names, locations, expressions of time and dates, organizations and more in a given text. Our approach "Octopub" does not aim to find new ways to improve named entity recognition process, rather than that it’s about finding a new, and yet smart way, to use NER in a way that we can extract sentiments of millions of people using Social Networks as a limitless information source, and Marketing for product promotion as the main domain of application.

Keywords: textmining, named entity recognition(NER), sentiment analysis, social media networks (SN, SMN), business intelligence(BI), marketing

Procedia PDF Downloads 558

Authors: Belkacem Soundes, Guezouli Larbi

Abstract:

This paper presents a robust and accurate method for text detection and localization over lecture videos. Frame regions are classified into text or background based on visual feature analysis. However, lecture video shows significant degradation mainly related to acquisition conditions, camera motion and environmental changes resulting in low quality videos. Hence, affecting feature extraction and description efficiency. Moreover, traditional text detection methods cannot be directly applied to lecture videos. Therefore, robust feature extraction methods dedicated to this specific video genre are required for robust and accurate text detection and extraction. Method consists of a three-step process: Slide region detection and segmentation; Feature extraction and non-text filtering. For robust and effective features extraction moment functions are used. Two distinct types of moments are used: orthogonal and non-orthogonal. For orthogonal Zernike Moments, both Pseudo Zernike moments are used, whereas for non-orthogonal ones Hu moments are used. Expressivity and description efficiency are given and discussed. Proposed approach shows that in general, orthogonal moments show high accuracy in comparison to the non-orthogonal one. Pseudo Zernike moments are more effective than Zernike with better computation time.

Keywords: text detection, text localization, lecture videos, pseudo zernike moments

Procedia PDF Downloads 122

Authors: Hend Alrasheed

Abstract:

Due to their different properties that have led to avoid several limitations of classic client/server systems, there has been a great interest in the development and the improvement of different peer to peer systems. Understanding the properties of complex peer to peer networks is essential for their future improvements. It was shown that the performances of peer to peer protocols are directly related to their underlying topologies. Therefore, multiple efforts have analyzed the topologies of different peer to peer systems. This study presents an overview of major findings of close experimental analyses to different topologies of three unstructured peer to peer systems: BitTorrent, Gnutella, and FreeNet.

Keywords: peer to peer networks, network topology, graph diameter, clustering coefficient, small-world property, random graph, degree distribution

Procedia PDF Downloads 354

Authors: Tatiana Stanchuliak

Abstract:

It is common knowledge that intonation is not laid over a ready text. On the contrary, intonation forms and accompanies the text on the level of its birth in the speaker’s mind. As a result, intonation plays one of the fundamental roles in the process of transferring a thought into external speech. Intonation structure can highlight the semantic significance of textual elements and become a ranging mark in understanding the information structure of the text. Intonation functions by means of prosodic characteristics, one of which is intonation salience, whose function in texts results in making some textual elements more prominent than others. This function of intonation, therefore, performs as organizing. It helps to form the frame of key elements of the text. The study under consideration made an attempt to look into the inner nature of salience and create a sort of a text intonation model. This general goal brought to some more specific intermediate results. First, there were established degrees of salience on the level of the smallest semantic element - intonation group, as well as prosodic means of creating salience, were examined. Second, the most frequent combinations of prosodic means made it possible to distinguish patterns of salience, which then became constituent elements of a text intonation model. Third, the analysis of the predicate structure allowed to divide the whole text into smaller parts, or units, which performed a specific function in the developing of the general communicative intention. It appeared that such units can be found in any text and they have common characteristics of their intonation arrangement. These findings are certainly very important both for the theory of intonation and their practical application.

Keywords: accentuation , inner speech, intention, intonation, intonation functions, models, patterns, predicate, salience, semantics, sentence stress, text

Procedia PDF Downloads 234

Authors: Hannah Kolano, Paul Nadan, Jeremy Ryan, Sophia Nielsen

Abstract:

The hooves of goats and other ruminants, or the family Ruminantia, are uniquely structured to adapt to rough terrain. Their hooves possess a hard outer shell and a soft interior that allow them to both conform to uneven surfaces and hook onto prominent features. In an effort to apply this unique mechanism to a robotics context, artificial feet for a hexapedal robot have been designed based on the hooves of ruminants to improve the robot’s ability to traverse unstructured environments such as those found on a rocky planet or asteroid, as well as in earth-based environments such as rubble, caves, and mountainous regions. The feet were manufactured using a combination of 3D printing and polyurethane casting techniques and attached to a commercially available hexapedal robot. The robot was programmed with a terrain-adaptive gait and proved capable of traversing a variety of uneven surfaces and inclines. This development of more adaptable robotic feet allows legged robots to operate in a wider range of environments and expands their possible applications.

Keywords: biomimicry, legged locomotion, robotic foot design, ruminant feet, unstructured terrain navigation

Procedia PDF Downloads 99

Authors: Ilia Zharikov, Philipp Nikitin, Ilia Vasiliev, Vladimir Dokholyan

Abstract:

With the increasing popularity of document analysis and recognition systems, text detection (TD) and optical character recognition (OCR) in document images become challenging tasks. However, according to our best knowledge, no publicly available datasets for these particular problems exist. In this paper, we introduce a Distorted Document Images dataset (DDI-100) and provide a detailed analysis of the DDI-100 in its current state. To create the dataset we collected 7000 unique document pages, and extend it by applying different types of distortions and geometric transformations. In total, DDI-100 contains more than 100,000 document images together with binary text masks, text and character locations in terms of bounding boxes. We also present an analysis of several state-of-the-art TD and OCR approaches on the presented dataset. Lastly, we demonstrate the usefulness of DDI-100 to improve accuracy and stability of the considered TD and OCR models.

Keywords: document analysis, open dataset, optical character recognition, text detection

Procedia PDF Downloads 138

Authors: Thelma Palaoag, Melanie Basa, Jezreel Mark Panilo

Abstract:

Checking short-answer questions and essays, whether it may be paper or electronic in form, is a tiring and tedious task for teachers. Evaluating a student’s output require wide array of domains. Scoring the work is often a critical task. Several attempts in the past few years to create an automated writing assessment software but only have received negative results from teachers and students alike due to unreliability in scoring, does not provide feedback and others. The study aims to create an application that will be able to check short-answer questions which incorporate information extraction. Information extraction is a subfield of Natural Language Processing (NLP) where a chunk of text (technically known as unstructured text) is being broken down to gather necessary bits of data and/or keywords (structured text) to be further analyzed or rather be utilized by query tools. The proposed system shall be able to extract keywords or phrases from the individual’s answers to match it into a corpora of words (as defined by the instructor), which shall be the basis of evaluation of the individual’s answer. The proposed system shall also enable the teacher to provide feedback and re-evaluate the output of the student for some writing elements in which the computer cannot fully evaluate such as creativity and logic. Teachers can formulate, design, and check short answer questions efficiently by defining keywords or phrases as parameters by assigning weights for checking answers. With the proposed system, teacher’s time in checking and evaluating students output shall be lessened, thus, making the teacher more productive and easier.

Keywords: information extraction, short-answer question, natural language processing, application

Procedia PDF Downloads 405

Authors: Bharatendra Rai

Abstract:

The sequence of words in text data has long-term dependencies and is known to suffer from vanishing gradient problems when developing deep learning models. Although recurrent networks such as long short-term memory networks help to overcome this problem, achieving high text classification performance is a challenging problem. Convolutional recurrent networks that combine the advantages of long short-term memory networks and convolutional neural networks can be useful for text classification performance improvements. However, arriving at suitable hyperparameter values for convolutional recurrent networks is still a challenging task where fitting a model requires significant computing resources. This paper illustrates the advantages of using convolutional recurrent networks for text classification with the help of statistically planned computer experiments for hyperparameter tuning.

Keywords: long short-term memory networks, convolutional recurrent networks, text classification, hyperparameter tuning, Tukey honest significant differences

Procedia PDF Downloads 86

Authors: Usama Shahid

Abstract:

Be it written documents, news columns, or students' essays, verifying the content can be a time-consuming task. Apart from the spelling and grammar mistakes, the proofreader is also supposed to verify whether the content included in the essay or document is relevant or not. The irrelevant content in any document or essay is referred to as off-topic text and in this paper, we will address the problem of off-topic text detection from a document using machine learning techniques. Our study aims to identify the off-topic content from a document using Echo state network model and we will also compare data with other models. The previous study uses Convolutional Neural Networks and TFIDF to detect off-topic text. We will rearrange the existing datasets and take new classifiers along with new word embeddings and implement them on existing and new datasets in order to compare the results with the previously existing CNN model.

Keywords: off topic, text detection, eco state network, machine learning

Procedia PDF Downloads 53

Authors: Wided Bakari, Patrice Bellot, Omar Trigui, Mahmoud Neji

Abstract:

This article constitutes an opening to think of the modeling and analysis of Arabic texts in the context of a question-answer system. It is a question of exceeding the traditional approaches focused on morphosyntactic approaches. Furthermore, we present a new approach that analyze a text in order to extract correct answers then transform it to logical predicates. In addition, we would like to represent different levels of information within a text to answer a question and choose an answer among several proposed. To do so, we transform both the question and the text into logical forms. Then, we try to recognize all entailment between them. The results of recognizing the entailment are a set of text sentences that can implicate the user’s question. Our work is now concentrated on an implementation step in order to develop a system of question-answering in Arabic using techniques to recognize textual implications. In this context, the extraction of text features (keywords, named entities, and relationships that link them) is actually considered the first step in our process of text modeling. The second one is the use of techniques of textual implication that relies on the notion of inference and logic representation to extract candidate answers. The last step is the extraction and selection of the desired answer.

Keywords: NLP, Arabic language, question-answering, recognition text entailment, logic forms

Procedia PDF Downloads 310

Authors: Mark K. Johnson

Abstract:

Arvo Pärt has composed over 80 text-based compositions based on nine different languages. But prior to 2015, it was not publicly known what texts the composer used in composing a number of his non-vocal works, nor the language of those texts. Because of this lack of information, few if any musical scholars have illustrated in any detail how textual structure applies to any of Pärt’s instrumental compositions. However, in early 2015, the Arvo Pärt Centre in Estonia published In Principio, a compendium of the texts Pärt has used to derive many of the parameters of his text-based compositions. This paper provides the first detailed analysis of the relationship between structural aspects of the Church Slavonic Eastern Orthodox text of Psalm 112 and the musical parameters that Pärt used when composing the string orchestral work Psalom. It demonstrates that Pärt’s text-based compositions are carefully crafted works, and that evidence of the presence of the ‘invisible’ hand of the composer can be found within every aspect of the underpinning structures, at the more elaborate middle ground level, and even within surface aspects of these works. Based on the analysis of Psalom, it is evident that the text Pärt selected for Psalom informed many of his decisions regarding the musical structures, parameters and processes that he deployed in composing this non-vocal text-based work. Many of these composerly decisions in relation to these various aspects cannot be fathomed without access to, and an understanding of, the text associated with the work.

Keywords: Arvo Pärt, minimalism, psalom, text-based process music

Procedia PDF Downloads 199

Authors: Thomas Ousterhout

Abstract:

The purpose of this study was to investigate if incorporating meaningful pictures of gestures and facial expressions in short sentences of text could supplement the text with enough semantic information to produce and N400 effect when probe words incongruent to the picture were subsequently presented. Event-related potentials (ERPs) were recorded from a 14-channel commercial grade EEG headset while subjects performed congruent/incongruent reaction time discrimination tasks. Since pictures of meaningful gestures have been shown to be semantically processed in the brain in a similar manner as words are, it is believed that pictures will add supplementary information to text just as the inclusion of their equivalent synonymous word would. The hypothesis is that when subjects read the text/picture mixed sentences, they will process the images and words just like in face-to-face communication and therefore probe words incongruent to the image will produce an N400.

Keywords: EEG, ERP, N400, semantics, congruency, facilitation, Emotiv

Procedia PDF Downloads 236

Authors: D. Zhigulskaya, V. Anisimov, A. Pikunov, K. Babanova, S. Zuev, A. Latyshkova, K. Сhernozatonskiy, A. Revazov

Abstract:

Electrophysiology of information processing in reading is certainly a popular research topic. Natural reading, however, has been relatively poorly studied, despite having broad potential applications for learning and education. In the current study, we explore the relationship between text categories and spontaneous electroencephalogram (EEG) while reading. Thirty healthy volunteers (mean age 26,68 ± 1,84) participated in this study. 15 Russian-language texts were used as stimuli. The first text was used for practice and was excluded from the final analysis. The remaining 14 were opposite pairs of texts in one of 7 categories, the most important of which were: interesting/boring, fiction/non-fiction, free reading/reading with an instruction, reading a text/reading a pseudo text (consisting of strings of letters that formed meaningless words). Participants had to read the texts sequentially on an Apple iPad Pro. EEG was recorded from 12 electrodes simultaneously with eye movement data via ARKit Technology by Apple. EEG spectral amplitude was analyzed in Fz for theta-band (4-8 Hz) and in C3, C4, P3, and P4 for alpha-band (8-14 Hz) using the Friedman test. We found that reading an interesting text was accompanied by an increase in theta spectral amplitude in Fz compared to reading a boring text (3,87 µV ± 0,12 and 3,67 µV ± 0,11, respectively). When instructions are given for reading, we see less alpha activity than during free reading of the same text (3,34 µV ± 0,20 and 3,73 µV ± 0,28, respectively, for C4 as the most representative channel). The non-fiction text elicited less activity in the alpha band (C4: 3,60 µV ± 0,25) than the fiction text (C4: 3,66 µV ± 0,26). A significant difference in alpha spectral amplitude was also observed between the regular text (C4: 3,64 µV ± 0,29) and the pseudo text (C4: 3,38 µV ± 0,22). These results suggest that some brain activity we see on EEG is sensitive to particular features of the text. We propose that changes in theta and alpha bands during reading may serve as electrophysiological tools for assessing the reader’s cognitive state as well as his or her attitude to the text and the perceived information. These physiological markers have prospective practical value for developing technological solutions and biofeedback systems for reading in particular and for education in general.

Keywords: EEG, natural reading, reader's cognitive state, theta-rhythm, alpha-rhythm

Procedia PDF Downloads 57