Ontologies have gained a lot of popularity and recognition in the semantic web because of their extensive use in Internet-based applications. Ontologies are often considered a fine source of semantics and interoperability in all artificially smart systems. Exponential increase in unstructured data on the web has made automated acquisition of ontology from unstructured text a most prominent research area. Several methodologies exploiting numerous techniques of various fields (machine learning, text mining, knowledge representation and reasoning, information retrieval and natural language processing) are being proposed to bring some level of automation in the process of ontology acquisition from unstructured text. This paper describes the process of ontology learning and further classification of ontology learning techniques into three classes (linguistics, statistical and logical) and discusses many algorithms under each category. This paper also explores ontology evaluation techniques by highlighting their pros and cons. Moreover, it describes the scope and use of ontology learning in several industries. Finally, the paper discusses challenges of ontology learning along with their corresponding future directions.

Web contains a vast amount of data which is accumulated, studied and utilized by a huge number of users on daily basis. A substantial amount of data on web is available in an unstructured format such as web pages, books, journals, and files. Acquiring appropriate information from such humongous data has become quite challenging and time consuming task. Trivial keyword based information retrieval systems highly depend on the statistics of data, thus face word mismatch problem due to inevitable semantic and context variations of certain word. Therefore, this marks the desperate need to organize such massive data into a structured format so that information can be easily processed in a large context by taking data semantics into account. Ontologies are not only being extensively employed in semantic web to store unstructured information in an organized and structured way but it has also raised the performance of diverse information retrieval approaches to a great extent. Ontological information retrieval systems retrieve data based on the similarity of semantics between user query and indexed data. This paper reviews modern ontology based information retrieval methods for textual, multimedia and cross lingual data types. Furthermore, we compare and categorize most recent approaches used in the above mentioned information retrieval methods along with their major drawbacks and advantages.

In order to accelerate the performance of various Natural Language Processing tasks for Roman Urdu, this article for the very first time provides 3 neural word embeddings prepared using most widely used approaches namely Word2vec, FastText, and Glove. The integrity of generated neural word embeddings is evaluated using intrinsic and extrinsic evaluation approaches. Considering the lack of publicly available benchmark datasets, it provides a first-ever Roman Urdu public dataset which consists of 3241 sentiments annotated against positive, negative, and neutral classes. To provide benchmark baseline performance over the presented dataset for Roman Urdu sentiment analysis, we adapt diverse machine learning (Support Vector Machine, Logistic Regression, Naive Bayes), deep learning (convolutional neural network, recurrent neural network), and hybrid deep learning approaches. Performance impact of generated neural word embeddings based representation is compared with other most widely used bag of words based feature representation approaches using diverse machine and deep learning classifiers. In order to improve the performance of Roman Urdu sentiment analysis, it proposes a novel precisely extreme multi-channel hybrid methodology which makes use of convolutional and recurrent neural networks along with pre-trained neural word embeddings. The proposed hybrid approach outperforms adapted machine learning approaches by the significant figure of 9% and deep learning approaches by the figure of 4% in terms of F1-score.

In order to provide benchmark performance for Urdu text document classification, the contribution of this paper is manifold. First, it provides a publicly available benchmark dataset manually tagged against 6 classes. Second, it investigates the performance impact of traditional machine learning-based Urdu text document classification methodologies by embedding 10 filter-based feature selection algorithms which have been widely used for other languages. Third, for the very first time, it assesses the performance of various deep learning-based methodologies for Urdu text document classification. In this regard, for experimentation, we adapt 10 deep learning classification methodologies which have produced best performance figures for English text classification. Fourth, it also investigates the performance impact of transfer learning by utilizing Bidirectional Encoder Representations from Transformers approach for Urdu language. Fifth, it evaluates the integrity of a hybrid approach which combines traditional machine learning-based feature engineering and deep learning-based automated feature engineering. Experimental results show that feature selection approach named as normalized difference measure along with support vector machine outshines state-of-the-art performance on two closed source benchmark datasets CLE Urdu Digest 1000k, and CLE Urdu Digest 1Million with a significant margin of 32% and 13%, respectively. Across all three datasets, normalized difference measure outperforms other filter-based feature selection algorithms as it significantly uplifts the performance of all adopted machine learning, deep learning, and hybrid approaches. The source code and presented dataset are available at Github repository https://github.com/minixain/Urdu-Text-Classification.

• A graph based novel approach for transforming RNA sequences into statistical vectors. • Development of a robust and precise deep learning classifier. • Utilization of attention mechanism capable of generating comprehensive feature space. • Incorporation of explainability suitable for highlighting distributions of nucleotides. • Development of a public web server use-able for predicting subcellular localizations.