JAWAHARLAL NEHRU TECHNOLOGICAL UNIVERSITY HYDERABAD MCA 2023 INFORMATION RETRIEVAL SYSTEMS Previous Question Paper (March/April - 2023 - R20)

Precision is a measure of the exactness or quality of an information retrieval system. It is defined as the ratio of the number of relevant documents retrieved to the total number of documents retrieved by the system. It indicates what proportion of the search results are actually relevant to the user's query.

Recall is a measure of the completeness or coverage of an information retrieval system. It is defined as the ratio of the number of relevant documents retrieved to the total number of relevant documents present in the entire collection. It indicates the system's ability to find all relevant information available.

In practice, there is often an inverse relationship between these two metrics. Increasing recall by broadening search criteria typically results in a decrease in precision, as more non-relevant documents are likely to be included in the retrieval set.

An N-gram is a contiguous sequence of n items from a given sample of text or speech. These items can be phonemes, syllables, letters, words, or base pairs. In information retrieval and natural language processing, N-grams are primarily used for tokenization, language identification, and approximate string matching.

The value of n defines the type of N-gram:

• Unigram (n=1): A single token.
• Bigram (n=2): A sequence of two adjacent tokens.
• Trigram (n=3): A sequence of three adjacent tokens.

N-grams are significant because they capture local context and word order, which simple bag-of-words models ignore. They are widely used in:

• Spelling Correction: Identifying common character sequences to suggest corrections.
• Search Engines: Handling phrases and multi-word queries by indexing bigrams or trigrams to improve precision.
• Language Modeling: Predicting the next word in a sequence based on the probability of the preceding n-1 words.

While larger values of n provide more context, they lead to data sparsity and an exponential increase in the size of the vocabulary or index, necessitating a balance between context depth and computational efficiency.