Utilizing machine learning to make smart investment choices in the California housing market.
- Perform clustering using k-means, GMM, and Fractal analysis on the real estate dataset to identify market segments.
- Explore methods for classifying and amalgamating different real estate datasets (scraping data) to create a more comprehensive dataset for analysis.
- Implement the Muller loop for classification to develop predictive models for real estate investment decisions.
- Conduct feature importance analysis using the random forest classifier to identify the most critical features in real estate investment models.
- Investigate the effects of upsampling and downsampling on key features to address data imbalance within the dataset.
- Apply linear regression models and the Muller loop for regression to predict future house prices for the next 1, 2, and 5 years based on historical data and influential factors.
- NLP techniques were employed to compare nominated and non-nominated poems against the Pushcart Prize benchmark, serving as a gold standard for poetry evaluation.
- The effects of Part-of-Speech (POS) substitutions in poems by William Shakespeare and William Wordsworth using NLP methods, shedding light on how these alterations influence semantic structure and interpretation in poetry.
- Fine Tuning a Mistral 7B large language model on Shakespeare's poetry to generate new works in his style and themes. Another integrated a Weaviate vector database with a local LLM to delve into the works and historical context of William Shakespeare for nuanced NLP tasks. Lastly, constructing and analyzing Knowledge Graphs derived from Shakespeare's poetry and related historical sources, exploring their structural, content, and thematic connections.