تاریخ طبری کے ضعیف اورجھوٹے راویوں کی روایات کا تحقیقی جائزہ

History of Tibari is the well-known book of late ‘allama ibne jar┘r ║ibar┘. Its real name is Tar┘kh- ul ’ummam wal Mul┴k. History of ║ibar┘ is considered the comprehensive and encyclopedia for the first three decades and the backbone in the history of Islam. He is considered a great and lofty character especially in the history of Islam, although all the historians of the present as well as of the past take guidance from his book. Inspite of the facts there are also baseless and false quotations written about Su╒┐ba’ kir┐m, explanation of which is not reasonable. As there are present some false, man-made and illogical sayings in Tar┘kh ║ibar┘. Therefore, an explanatory summary is presented of the narrators so that it may be clyster clear that ‘Allama ║ibar┘ is trusty and worthy but his works are the combination of both facts and false.

Polarity Estimation and Aggregation in Aspect Based Sentiment Analysis

Sentiment analysis and opinion mining (OM) is a developed part of the research which evaluates people‟s views, ideas or sentiments. Sentiment analysis plays a key role in a classification task because a bulk of contents is generated and issued on the Internet per day. Sentiment lexicons have been used effectively to categorize the sentiment of user review corpuses. According to research perspective, sentiment analysis floats in three different directions i.e. document level, sentence level and aspect level. Aspect Based Sentiment Analysis (ABSA) deals with the exploration of feelings, opinions, facts and emotions in the phrases which are expressed by the humans in a particular review. It allows user to identify the feelings and attitudes of a particular person or people by analyzing comments, Tweets, blogs and reviews about all the aspects. In most research techniques, the ABSA process involves classification of user reviews into three classes i.e. positive, negative or neutral from textual dataset of reviews. Such classification of the sentiment is called sentiment polarity. In today''s research sentiment polarity can be consider as one of the major task in Opinion Mining. Most common techniques in practice for polarity estimation attempt to identify the main i.e. the most commonly and frequently deliberated features of the entity e.g., „screen‟, „memory‟, ''battery'' of a particular mobile brand and to compute the mean polarity of the review per feature like how much positive, negative or may be neutral the ideas are on average for each feature. Most of such techniques are lexicon or corpus based which is domain specific. The machine learning technique is remarkable, but it divides the sentiment polarity into three categories i.e. positive, negative, or neutral based on some training data. Moreover, such techniques usually Asif Nawaz, Reg. No. 109-FBAS/PHDCS /F14vii fail to concern with the fuzziness of sentiment polarity and polarity intensity of the sentiment words. This research proposes a new technique for polarity estimation and aggregation, the whole method consists of three main subtasks, the first task is the aspect extraction which extracts the core features of the entity being deliberated based on the similarity measure like NGD and ConceptNet. Polarity estimation task will calculate the polarity with respect to each aspect using max entropy model which further leads to some rela time applications for event identification. Finally polarity aggregation will aggregate the overall polarity to identify that how much an individual review is close to a particular class. Furthermore, the proposed conceptual framework is applied on various domains which are in research trend like Medical and Software Specific Word Repositories. From the evaluation of experimental results, it is concluded that performance of the proposed conceptual framework is explicitly up to the mark. Keywords: ABSA, Aspect Identification, Polarity Estimation, Event Identification, FCA, NGD.