Leon Coe’s Post

Unveiling Emotions: Harnessing SVM in R Studio for Social Media Sentiment Analysis. Explore the spectrum of emotions from anger to happiness, smiles to worries, as we employ SVM modeling to decode sentiments in social media comments. #EmotionAnalysis #SVM #RStudio https://lnkd.in/gMj-ZRBD

TASK 3: 🚀 Excited to share my latest project where I delved into customer purchase behavior prediction using a decision tree classifier! 🌳💼 📊 Leveraging a dataset from the banking sector, I explored demographic and behavioral data to predict whether a customer will purchase a product or service. 🛠 I preprocessed the dataset, encoded categorical variables, and split it into training and testing sets. Using scikit-learn, I trained a decision tree classifier and evaluated its accuracy. Prodigy InfoTech https://lnkd.in/g2yjQx6h

Progress Update: Completed Session 3 of the Machine Learning Zoomcamp! Worked on a Churn Prediction Project and covered the following: - Data preparation and validation framework setup - Exploratory Data Analysis (EDA) - Feature importance using churn rate, risk ratio, mutual information, and correlation - One-hot encoding for categorical features - Logistic regression training with Scikit-Learn - Model interpretation and practical usage Big thanks to Alexey Grigorev and the DataTalksClub community for their support! 🙌 Check out my progress and the code here: https://lnkd.in/djemp7Jf #MachineLearning #DataScience #Classification #DataPreparation #FeatureEngineering #DataTalksClub #ContinuousLearning

Jerry Liu's innovative approach to combining LlamaParse's parsing capabilities with knowledge graphs to answer complex queries. By constructing a knowledge graph from #LlamaParse's parsing output, Liu develops a Retrieval-Augmented Generation (RAG) pipeline. This pipeline uses vector search to retrieve initial nodes, followed by graph traversal to find related nodes, ultimately enabling an agent to answer complex queries #Effectively!.

𝗔𝗱𝘃𝗮𝗻𝗰𝗲𝗱 𝗥𝗔𝗚𝘀 by Hand ✍️ ~ Download https://by-hand.ai/rag Thank you all for joining my webinar on "Beginner's Guide on RAG" yesterday. I was really surprised by the huge turnout! Thank you! 🙌 I wanted to share an animated version of the last section of my webinar, where I gave an overview of several advanced RAG techniques. • Multiple Embedding Spaces • RankGPT • Multi-query Retrieval • Contextual Compression • Hypothetical Document Embeddings If you are interested in how I explained these techniques, I encourage you to watch the recording. There are many more advanced topics to talk about on RAG. I may do more lectures later. Thanks for hosting me, team SingleStore! Matt Brown Akmal Chaudhri, Esq. Yukthi Dwivedi Thanks for promoting this webinar! Aishwarya Srinivasan Alex Wang Brij kishore Pandey Aman Chadha Thanks for showing me cool Superlinked demos to help prepare this webinar! Daniel Svonava Mór Kapronczay Thanks for answering a bunch of technical questions about LlamaIndex! Ravi Theja Desetty Val Andrei Fajardo #rag #vectordatabases #aibyhand If you find this visual helpful, please [REPOST ♻️] to share this with others.

#TipTuesday | Choosing a chunk size for Retrieval Augmented Generation (RAG). A key step in the RAG pipeline is chunking your unstructured source data - dividing a large body of information into manageable pieces. These chunks will be searched through for relevance to the user query and used as the context for a response. They must enable efficient search operations while containing meaningful information. Too large and it may contain information on multiple topics, reducing its similarity score when compared to a single one of those topics in a query. Too small and it may not contain a complete or useful factual statement about a given subject. When deciding on an appropriate chunk size, you should start by understanding the source data: - How it’s structured - How much there is - How verbose it is For example, a bullet point list will likely be relatively information-dense and suited to a smaller chunk size than a description given in a conversational style. Take a look at our blog for more details: https://lnkd.in/eV8pjh8k #RAG #RetrievalAugmentedGeneration #SourceData 

After finishing the Machine Learning Specialization, I applied the Regresison method (Linear, Lasso and Ridge Regresison) on the Life Expectancy dataset In this project, I needed to get some insights of what affects Life Expectancy in different countries, and in order to do so: 1- Have a quick look at the correlation between features using a heatmap 2- The Status of provided countries, by creating a pie chart to express the precentage of developed to developing countries 3- Check the population of different countries by creating a box plot 4- Check the Life Expectancy range through the years 5- Get a better insight of the spreading of diseases provided in the dataset in different countries 6- Check the correlation between Life Expectancy and spreading Diseases in different Countries 7- Train the model using different Regression techniques to choose the best model with highest accuracy notebook link on GitHub: https://lnkd.in/dTrT6-pn notebook link on Kaggle: https://lnkd.in/dKqQaqWa

Useful one explaining the basics, thanks Tom Yeh !!

LlamaParse and Knowledge Graphs, two great tastes that taste great together! In this notebook Jerry Liu ⭐️ Uses first-class parsing from LlamaParse as the raw material to construct a knowledge graph ⭐️ Builds a RAG pipeline on top of the graph, retrieving initial nodes via vector search then related nodes via graph traversal ⭐️ Builds an agent on top of the RAG pipeline to answer complex queries Check it out here: https://lnkd.in/gjk67KVz

🚢 Exploring the Titanic Dataset: Insights and Analysis 📊 Over the past week, I dived deep into the Titanic dataset, a classic dataset often used for data science and machine learning projects. Task 2: Perform Data Cleaning and EDA on a dataset of your choice, such as the titanic dataset from Kaggle. Explore the relationship between variables and identify patterns and trends in the data. Here's a summary of what I did and what I learned: 🔍 What I Did: 1. Data Generation: Created a synthetic Titanic dataset, simulating various passenger details including age, fare, passenger class, and survival status. 2. Data Cleaning: Handled missing values, encoded categorical variables, and scaled numerical features to prepare the dataset for analysis. 3. Exploratory Data Analysis (EDA):  - Analyzed survival distribution and investigated survival rates based on gender, passenger class, and port of embarkation.  - Explored the age and fare distributions to understand the socio-economic status of passengers.  - Examined family size and its impact on survival rates. 4. Correlation Analysis: Computed and visualized the correlation matrix to identify relationships between numerical variables. 📈 What I Learned: - The critical importance of data cleaning and preparation before any analysis. - Visualization techniques to uncover patterns and trends in the data. - The significant impact of socio-economic status (passenger class) and gender on survival rates. - The usefulness of correlation analysis in identifying relationships between variables. You can check out the detailed analysis and the code on my GitHub repository: https://lnkd.in/dKaUws9M This project has been an excellent opportunity to apply data science techniques and derive meaningful insights from data. I'm excited to continue exploring and learning in the field of data science! #DataScience #MachineLearning #EDA #Python #TitanicDataset #GitHub #LearningJourney #DataAnalysis #ProdigyInfoTech