Search or add a thesis

Advanced Search (Beta)
Home > Biomedical Data Retrieval Using Enhanced Query Expansion

Biomedical Data Retrieval Using Enhanced Query Expansion

Thesis Info

Access Option

External Link

Author

Muhammad Qadeer

Institute

Virtual University of Pakistan

Institute Type

Public

City

Lahore

Province

Punjab

Country

Pakistan

Thesis Completing Year

2019

Thesis Completion Status

Completed

Subject

Software Engineering

Language

English

Link

http://vspace.vu.edu.pk/detail.aspx?id=297

Added

2021-02-17 19:49:13

Modified

2024-03-24 20:25:49

ARI ID

1676721017310

Similar


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.
Loading...
Loading...

Similar Books

Loading...

Similar Chapters

Loading...

Similar News

Loading...

Similar Articles

Loading...

Similar Article Headings

Loading...

یہ ہنر آزما کے دیکھتے ہیں

یہ ہنر آزما کے دیکھتے ہیں
اب کے اُن کو بھلا کے دیکھتے ہیں

رخ سے سرکے کبھی جو آنچل، ہم
معجزے اُس خدا کے دیکھتے ہیں

اب کے ہم بھی تو بے وفا ہو کر
جوش اُس کی وفا کے دیکھتے ہیں

کوئی اپنا نکل بھی سکتا ہے
ہاتھ سب سے ملا کے دیکھتے ہیں

درد میں آئے کچھ کمی شاید
آس کا گھر جلا کے دیکھتے ہیں

مل ہی جائے گا شعلۂ امید
گردِ اِیذا ہٹا کے دیکھتے ہیں

یاروں کی کھوجتے ہیں اب نیت
اپنا مقتل سجا کے دیکھتے ہیں

جو بھی ہوتا ہے خوش یہاں اُس کو
درد سے، منہ اُٹھا کے دیکھتے ہیں

جو بھی آتا ہے اُس گلی سے ہم
پاس اُس کو بٹھا کے دیکھتے ہیں

Anxiety Disorders in Pregnant Women: Causes and How to Overcome Them

The article explains the origins of anxiety disorders during pregnancy, and their treatment. Pregnant women's anxiety is normally manifests itself as emotions, perceptions, and behaviors, which is mediated by biological, social, and genetic influences. Probability analysis is done of this kind. Using a survey of 1500 rural residents as a random number between 100 and 1600 produces a 200% response rate. Those findings revealed that 60% of the participants had adequate reproductive age and 51% had mild to moderate anxiety, and 78.4% had high anxiety. The reasons leading to pregnant women's anxiety include their age, working status, lack of care from a boyfriend, having had previously given birth, and their wellbeing before becoming pregnant, all of the woman's family members, and maternal well-being. To resolve this is by sufficient relaxation, anti-depressants, a safe diet, and physical exercise, along with learning how to behave more positively, and by means of ourselves, or through prantal massage.

Study of Pyocyanin Induced Virulence & Pathogenicity of Nosocomial Infectious Agent Pseudomonas Aeruginosa and its Prevention

Pseudomonas aeruginosa is a ubiquitously distributed and often, nosocomial infectious agent. It is identified as one of the top three nosocomial infectious agents, causing complicated and persistent infections. It has notorious infection feactures such as biofilm formation and antibiotic resistance, making it a significant healthcare challenge. P. aeruginosa is equipped with a diversified arsenal of virulence factors which help in infection establishment and progression against the host immune system. Among many other, P. aeruginosa produces Pyocyanin (PCN), a redox active blue phenazine pigment, considered to be its primary virulence factor. This blue pigment plays a critical role in infection establishment and survival of the pathogen. Studies have implicated PCN with multiple bioactive properties due to its redox active potential, i.e. immuno-modulation, pro-apoptotic, enzyme inactivation and others. The ability to diffuse through biological membranes imparts a new dimension for PCN mediated virulence of P. aeruginosa in the host body. So far PCN mediated virulence has been considered as localized incidence at the site of infection i.e. effect of PCN on the ciliary movement of epithelial lining or inactivation of DUOX1 enzyme in the respiratory system. The emergence of Pan-drug resistant or ‘Superbug strains’ at pandemic scale is a critical threat to health care setups. The ineffectiveness of antibiotics against P. aeruginosa puts the vulnerable groups at serious risk as the World Health Organization is foreseeing the dawn of Post-antibiotic era. Lack of preventive vaccines against P. aeruginosa is another challenge that puts vulnerable groups at greater risk. A comprehensive strategy was devised to better understand PCN mediated P. aeruginosa virulence. This dissertation reports the potential impact of PCN in Central Nervous System during P. aeruginosa infection. Cytotoxic impact of PCN was studied in murine models. Blood brain barrier permeability was assessed computationally and confirmed via HPLC. Murine behavioral tests demonstrated neurodegenerative and cognitive impairment potential of PCN. Systemic redox potential was evaluated by employing multiple in silico tools and glutathione-S-transferase assay, PCN-induced oxidative stress in brain, lungs and cardiac tissues. For dealing with PCN-mediated virulence, PCN biosynthesis was targeted via drug repurposing in a dual manner: 1) Study of PCN chemical-protein and drug drug interactions revealed Piperaquine as an anti-virulence therapeutic agent as it has an effective anti-PCN potential. While the structure based study demonstrated aspirin’s potential in inhibiting PCN, unlike ibuprofen. 2) Innate immune component DUOX1 and its homologue NOX4 capable of inactivation of PCN were computationally studied for their sequential, structural and functional aspects, and were annotated. These two schemes could contribute towards the development of Pseudomonas specific anti-virulence approaches independent of antibiotic resistance phenomenon. In a bigger scenario genomic data was explored for dealing with the antibiotic resistance challenge. Using reductive genomic screening, potential peptide vaccine candidates and prospective novel drug targets were predicted. This dissertation focuses on PCN mediated virulence of P. aeruginosa and strategies for prevention and management of infection.