M.Phil / MS
Permanent URI for this collectionhttps://repository.cuilahore.edu.pk/handle/123456789/36
This collection archives the complete set of theses produced by students of the COMSATS University Islamabad, Lahore Campus.
Browse
4 results
Search Results
Item Cross Genre Author Profiling Using Semantic Tagger Based Approach(Library Information Services, COMSATS University Islamabad, Lahore Campus, 2017) Humaira Muqdes; FA15-RCS-023; LHR TP 7000; Dr. Rao Muhammad Adeel NawabAuthor profiling is a task to predict one or more traits of an author (e.g. age, gender etc.), from his/her written text. The collaborative environments have significantly increased the chance of fake author profiles, particularly over the social media. The field of automatically detecting an author’s profile from written text has potential applications in marketing, forensics, security, detecting fake profiling and harassment cases. This research work aims to explore the problem of cross genre author profiling, in which training dataset is in one genre and test dataset is in another genre. We explored three different methods for cross-genre author profiling: (1) stylometry based approach, (2) content based approach, (3) semantic tags based approach and (4) combination of previous three approaches. As far as we are aware semantic tags based approach and combination of approaches has not been previously used for cross genre author profiling problem. Evaluation was carried out using four benchmark author profiling corpora: (1) PAN-AP-14 Social Media Corpus, (2) PAN-AP-14 Hotel Reviews corpus, (3) PAN-AP-14 Blogs Corpus and (4) PAN-AP-16 Twitter Corpus. Accuracy was used an evaluation measure. Above mentioned techniques were applied for both same and cross genre problems. For same genre, best results were obtained using content based technique on PAN-AP 14-Hotel Reviews Corpus (Accuracy =55.26) for Age (Accuracy= 66.59) for gender content based techniques for PAN-AP-Blogs Corpus (Accuracy = 67.57) for age, Semantic tagger based technique (Accuracy = 78.38) for gender. Content based technique on PAN-AP-14-social media Corpus (Accuracy =55.26) for Age (Accuracy= 78.95) for gender. Content based technique on PAN-AP-16-Twitter Corpus (Accuracy =54.76) for Age (Accuracy= 74.42) for gender. For Cross Genre Training Corpus PAN-14-Blogs and testing Corpus PAN-14-Hotel reviews for age (Accuracy = 56.96) for gender training Corpus is PAN-16-Twitter (Accuracy= 58.97). Content Based Technique for Training Corpus PAN-14-Hotel reviews and testing Corpus PAN-14-Blogs for age (Accuracy = 40.81) for gender Stylistic based for training Corpus is PAN-14-Blogs (Accuracy= 59.86). Content based technique for Training Corpus PAN-14-Social Media and testing Corpus PAN-14-Blogs for age (Accuracy = 40.81) for gender Content based technique training Corpus is PAN-14-Blogs (Accuracy= 59.86). Stylistic based technique for PAN-16- Twitter training corpus and testing corpus PAN-14-Blogs (Accuracy = 40.81) for age. Content based technique for gender where testing corpus is PAN-14-Blogs (61.22) for gender.Item Development of Large Scale English-Urdu Machine Translation Corpus for Statistical and Neural Machine Translation Systems(Library Information Services, COMSATS University Islamabad, Lahore Campus, 2019) Moodser Hussain; FA17-RCS-005; LHR TP 5623; Dr. Rao Muhammad Adeel NawabThe process of automatically converting the text of one natural language into an other is known as Machine Translation (MT). The two state-of-the-art machine translation techniques are Statistical Machine Translation (SMT) and Neural Ma chine Translation (NMT). In both MT techniques, a large-scale parallel aligned corpus of source & target language text is the most basic requirement to develop quality models. For the Urdu language, there are minimal parallel resources de veloped until now, and these existing resources are not enough to produce some quality translation systems. This research study is majorly focusing on: (1) the development of a large-scale semi-automatically aligned English-Urdu parallel corpus of more than 150k sen tences by collecting the data from different domains. (2) the compilation of large sized monolingual data for English and Urdu language by merging existing cor pora. (3) the development of a character-separated parallel corpus for supervised transliteration (by using Wikipedia titles) (4) the evaluation of impact generated by language model and monolingual corpus on translation quality (by using differ ent data and word-order in language models training). (5) development of Phrase based SMT (PB-SMT) and Recurrent Neural Network-based NMT (RNN-NMT) systems to evaluate and compare their performance on the proposed parallel cor pus. (6) the usage of supervised transliteration and ‘minimum error rate training’ to enhance the quality of SMT. To evaluate the performance of MT systems, ‘Bilingual Evaluation Understudy (BLEU)’ is used as an evaluation measure. The best-achieved scores for English to-Urdu translations are 37.39 and 32.70 by using PBSMT and RNN-NMT, re spectively. While for Urdu-to-English translations, PBSMT and RNN-NMT yield the best score of 28.20 and 30.63 respectivelyItem Toxic Comment Classification of Roman Urdu Text(Library Information Services, COMSATS University Islamabad, Lahore Campus, 2019) Waheed Abbas; FA17-RCS-013; LHR TP 5624; Dr. Rao Muhammad Adeel NawabWe are living in an era of technology where most of us have easy access to the inter net. Due to internet bloom, the use of social media and discussion forum, in particular for communication, increase tremendously in recent years. But this advancement also opens doors for trolls who poison these social media and forums by their abusive behav ior toward others. Detection of toxic language online is becoming an important issue in recent years. A toxic comment is defined as rude, unreasonable, or disrespectful com ments which likely to make you leave a conversation. Current methods for handling toxicity online are often heavily dependent on manual moderation, and not scalable enough to handle growing users day by day. In this thesis, we investigated the problem of toxic language detection for Roman Urdu (transliteration of Urdu in English) text because Roman Urdu is widely used to communicate online in Pakistani and Indian community. Various work has been done to detect toxicity in the English language, but due to unavailability of large Roman Urdu corpora, there is no work done so far in Roman Urdu. The aim of this thesis work is two folds: first is to develop large scale corpus for detection of toxic language in Roman Urdu, second is to apply different proven classical machine learning and state-of-the-art deep learning methods on the developed corpus. In the first step, we collected data from the YouTube comment section to develop Roman Urdu corpus. The final corpus consists of 10,863 unique comments where we have 5,887 toxic and 4,976 non-toxic comments. We performed manual annotation of corpus and got it reviewed from another annotator. We applied proven classical ma chine learning algorithms like Logistic Regression, K-Nearest Neighbors, Naive Bayes, Random Forest, Support Vector Machine and Multi-layer Perceptron along with state of-the-art deep learning methods like Long Short-Term Memory, Bidirectional Long Short-Term Memory, and Convolutional Neural Networks. Best results obtained from Random Forest with character 4-gram (Accuracy = 0.964, Precision = 0.980, Recall = 0.953, F1 = 0.966).Item Detecting Urdu Semantic Textual Similarity through Word and Sentence Embedding Techniques(Library Information Services, COMSATS University Islamabad, Lahore Campus, 2019) Muhammad Farhat Ullah; FA17-RCS-019; LHR TP 5625; Dr. Rao Muhammad Adeel NawabSemantic Textual Similarity (STS) evaluates the degree to which two sentences or short texts are semantically proportional to one another. STS is one of the significant prob lems in the field of Natural Language Processing (NLP). Text reuse and plagiarism detection are famous examples of STS. STS has three types; (1) Monolingual STS: if the source and suspicious short texts are in the same language. (2) Multilingual STS: if the source is in one language and sus picious short text is more than two languages. (3) Cross-lingual STS: if the source in one language and suspect is in the other language, often translation of each other. STS could be found several levels, for example, word, sentence, paragraph, and document level. Urdu is one of the low resource languages. It’s the National Language of Pakistan, also widely spoken and used in electronic, print media of Pakistan, India, and Bangladesh. The main aim of this thesis is to develop techniques that measure STS as paragraph level for the Urdu language. Our thesis aims to develop and investigate the new feature extraction techniques to ad dress the problem of STS for Urdu. We divide it into three layers, (1) Train Word and Sentence Embedding models on Urdu datasets, (2) Apply these new feature extraction techniques to extract feature from Urdu short text pairs, and (3) Apply machine learn ing classification algorithms for Urdu STS. In the first step we train Word and Sentence Embedding models on Urdu datasets. By using these embedding models, we extract the word, and sentence embedding features from pre-processed Urdu Short Text Reuse Corpus (USTRC) short texts. After that, we find the cosine similarity between these extracted feature vectors, then apply classification algorithms on similarity to classify short texts into verbatim, paraphrased and independently written. In word embedding techniques, we used Word2Vec, GloVe, and FastText with Addition, Average, and Mul tiplication Functions also we explored Smooth Inverse (SI) and Term Frequency (TF) weighted word embedding techniques. In sentence embedding techniques we have used two unsupervised (sent2vec and LASER) and two supervised (InferSent and BERT) ix techniques. We have used seven machine learning algorithms to classify similarity score including, Naive Bayes (NB), Support Vector Machine (SVM), Logistic Regression (LR), Random Forest (RF), K-Nearest Neighbour (KNN), Gradient Boost (GB), Multi-Layer Percep tron (MLP). To evaluate these classifiers, we apply F1 measure. We got best F1 measure = 0.68, 0.75, 0.92, 0.70 by using sent2vec sentence embedding technique with GB and MLP classifiers.