SPPU · Final Year Computer Engineering · 2019 Course (Revised)
Information Retrieval
410245A · Elective IV · Semester VII · 3 credits · In-semester 30 + end-semester 70
One of the Elective IV options. Shares notes with the third-year course of the same name; the final-year units add their own topics.
Unit 1: Introduction, Basic Techniques & Tokens
7 hoursThe IR system and its architecture; Boolean, TF-IDF, vector, probabilistic and latent semantic indexing models; tokenising, stop words and stemming.
Marks: 15 in the in-semester exam
- Information retrieval systems and search engine architecture
- Boolean retrieval
- TF-IDF and the inverted index
- Vector space model and cosine ranking
- Latent semantic indexing
- Text preprocessing: tokenisation, stemming and lemmatisation
Unit 2: Static Inverted Indices and Query Processing
7 hoursIndex components and life cycle; sort- and hash-based dictionaries; in-memory, sort-based, merge-based and disk-based construction; document-at-a-time and term-at-a-time ranked retrieval, precomputed scores, impact ordering, query optimisation; generalised concordance lists.
Marks: 15 in the in-semester exam
- Dictionaries and index construction
- Query processing: document-at-a-time and term-at-a-time
- Query optimisation and impact ordering
Unit 3: Index Compression and Dynamic Inverted Indices
7 hoursModelling and coding, Huffman and arithmetic coding, symbolwise compression; parametric and nonparametric gap compression, context-aware methods, decoding performance, document reordering; incremental updates, contiguous and noncontiguous lists, deletions and modifications.
Marks: 17–18 in the end-semester exam
- Text compression: Huffman and arithmetic coding
- Compressing postings lists
- Dynamic inverted indices
Unit 4: Probabilistic Retrieval, Language Modelling, Categorization & Filtering
7 hoursBinary independence model, BM25, relevance feedback, field weights; language models and smoothing, divergence from randomness, passage retrieval; categorisation and filtering with linear, similarity-based and probabilistic classifiers, generalised linear and information-theoretic models.
Marks: 17–18 in the end-semester exam
- Probabilistic retrieval and BM25
- Relevance feedback
- Language models for IR
- Text categorisation and filtering
Unit 5: Measuring Effectiveness and Measuring Efficiency
7 hoursTraditional effectiveness measures, TREC, statistics in evaluation, minimising adjudication effort, nontraditional measures; efficiency criteria, query scheduling, caching with Redis and Memcached.
Marks: 17–18 in the end-semester exam
- IR evaluation metrics
- TREC and statistical evaluation
- Measuring efficiency: query scheduling and caching
Unit 6: Parallel Information Retrieval, Web Search
7 hoursParallel query processing and MapReduce; the structure of the web, queries and users, static and dynamic ranking, evaluating web search; crawlers, Scrapy and BeautifulSoup.
Marks: 17–18 in the end-semester exam
- Parallel information retrieval and MapReduce
- Web search: challenges, architecture and ranking
- Web crawling and meta-search
- Web scraping with Python
Unit contents are summarised from the official SPPU syllabus . AiCanCode.org is not affiliated with Savitribai Phule Pune University.