فیوض الحرمین کے دو اردو تراجم کا تقابلی جائزہ

Fayyuz-ul-Herman is the book of Shah Waliullah. In which he describes the circumstances of the two years he spent in Macca and Madina. This book is in Arabic and there are two translations in Urdu. The name of the first Translation is Saadat e Konin. The second translation's name is Mushahidat o Maaraf. This article has a comparative study of both the merits and the shortcomings are described.

Biomedical Data Retrieval Using Enhanced Query Expansion

Biomedical data is growing up rapidly and a better retrieval system is the need for its utilization. A basic problem while retrieving data from a system related to the queries is mismatch of words, which indicates the use of dissimilar words for expressing the identical concepts in given queries and in the stored documents. Two techniques are commonly used to solve this problem i.e. query paraphrasing as well as query expansion. Query paraphrasing refers that the query is paraphrased by using synonyms of terms in the query. Query expansion techniques are further categorized as local and global. Local query expansion technique focuses on the analysis of the documents having top ranks retrieved for a query. Different ranking models have been introduced to rank documents in collections based on terms and features. A collection of candidate terms is obtained for expanding the given query from these documents. On feature selection from term pool, final selected candidate expansion terms contain a few terms which cause query drift problem. To overcome this problem the semantic filtering technique was used. Semantic similarity measures are the basic techniques for successful semantic filtering. However, global query expansion relies on the analysis of the whole collection to find out word relationships. Synonyms of query words are extracted from a dictionary or thesaurus. In this research, we evaluated the famous probability-based ranking models such as LM-Dirichlet, LM Jelinek Mercer and BM25 for biomedical data retrieval process. We performed experimental analysis using diverse preprocessing techniques iteratively on 36 biomedical related queries for the evaluation. State of the art biomedical data set Trec Genomic was used as a core for whole experimentation. It was observed that BM25 was the best information retrieval model for biomedical data. We used different terms scoring techniques such as Baseline, BNS, Chi-Square, Codice, BIM, KLD, LRF, PRF and RSV to score the terms related to the query. The average of MAP scores of all the queries was compared that exhibited BNS term scoring technique is the best for biomedical data. Different semantic similarity measures such as Path-based, Wu and palmer, Leacock and Chodorow were applied on terms extracted from BNS to get most appropriate terms for query expansion. Finally, queries expanded with the most similar terms each time and documents retrieved through the expanded queries and the MAP results were evaluated for the purpose of final declarations of this research. The results of biomedical data retrieval through query expansion were improved and the LCH semantic similarity measuring technique found best for query expansion in biomedical data retrieval system.