JAWAHARLAL NEHRU TECHNOLOGICAL UNIVERSITY HYDERABAD MCA 2021 SOCIAL MEDIA INTELLIGENCE Previous Question Paper (February - 2021 - R17)

Text mining, also known as text data mining or text analytics, is the automated process of deriving high-quality information, patterns, and insights from unstructured text data. Unlike traditional data mining which operates on structured databases, text mining transforms raw, natural language text into a structured format to facilitate analysis and decision-making.

The process typically involves several key stages:

1. Information Retrieval: Gathering relevant text documents from various sources such as social media, emails, or web pages.
2. Preprocessing: Cleaning the data through tokenization (splitting text into words), stop-word removal (deleting common words like 'the' or 'is'), stemming or lemmatization (reducing words to their root form), and handling noise like HTML tags or special characters.
3. Transformation: Converting the processed text into a numerical representation, such as the Bag-of-Words model or Term Frequency-Inverse Document Frequency (TF-IDF), which allows algorithms to process the text mathematically.
4. Pattern Discovery: Applying techniques like classification (categorizing text into predefined labels), clustering (grouping similar documents), sentiment analysis (determining emotional polarity), and entity extraction (identifying names, locations, or organizations).
5. Evaluation and Visualization: Interpreting the extracted patterns and presenting them through dashboards or reports to provide actionable intelligence.

Text mining is essential in social media intelligence for monitoring brand reputation, identifying emerging trends, detecting public sentiment, and analyzing user behavior across massive datasets that would be impossible to process manually.

Expectations act as a cognitive filter that shapes how individuals perceive, process, and interpret information, fundamentally influencing the formation and modification of opinions. When an individual holds a pre-existing expectation about a subject, they are prone to confirmation bias, where they selectively attend to information that aligns with their beliefs while disregarding contradictory evidence. This psychological phenomenon ensures that opinions are often reinforced rather than challenged by new data.

Expectations influence opinion through several mechanisms:

• Selective Perception: People notice details that confirm their existing mental models, leading to a distorted view of reality.
• Interpretation Bias: Ambiguous information is interpreted in a way that supports the expected outcome, effectively neutralizing potential counter-arguments.
• Anchoring Effect: Initial expectations serve as a reference point; subsequent information is judged relative to this anchor, making it difficult to shift an opinion significantly once it is formed.

For example, consider a consumer who expects a specific brand to be of low quality due to its low price point. When using the product, if they experience a minor technical glitch, they will immediately interpret it as definitive proof that the brand is unreliable, ignoring the fact that the product has performed well in other areas. Conversely, if they experience a positive feature, they may dismiss it as an anomaly or a temporary fluke. In this case, the consumer's initial expectation dictates their final opinion, causing them to maintain a negative stance regardless of the actual product performance.