Hello, I'm Anupam! 👋 I'm passionate about technology and have a positive outlook, always eager to grow and learn.
This case study aims to explore the key factors (or driver variables) influencing loan default, specifically identifying variables that strongly indicate default. The finance company can apply this insight to enhance its portfolio and risk assessment. The project employs techniques such as Exploratory Data Analysis, distribution plotting, visualization, and Hypothesis Testing to analyze the impact of consumer and loan attributes on default tendencies.
Python packages used Numpy, Pandas, Matplotlib and Seaborn
#Python, #EDA, #DataVisualization, #StatisticalInference, #HypothesisTesting
This case study is made to really look into why people want to use shared bikes and what are the factors that influence it.The goal is to assist BoomBikes , a USA based bike-sharing provider in developing a Multiple Linear Regression model, predicting variations in demand, and providing valuable insights for strategic business decisions to accelerate revenue growth.
Python packages used sklearn, statsmodels, Numpy, Pandas, Matplotlib and Seaborn
#SupervisedLearning, #LinearRegression, #ModelSelection, #FeatureSelection, #MinMaxScaler, #VarianceInflationFactor
This case study aims to understand and solve House Price Prediction problem within the Real Estate domain, exploring how data can be leveraged effectively to tackle business challenges like this. Our objective is to forecast house sales price in a specific area and understand which factors are responsible for higher property value. We aim to address this challenge by employing Advanced Regression Technique like Ridge and Lasso with Cross Validation.
Python packages used sklearn, statsmodels, Numpy, Pandas, Matplotlib and Seaborn
#SupervisedLearning, #RidgeRegression, #LassoRegression, #ModelSelection, #FeatureSelection, #CrossValidation, #HyperparametersTunning
In the telecommunications industry, customer retention is a significant challenge, with customers often switching operators in search of more attractive schemes and offers. This case study’s objective is to develop a machine learning model that can accurately predict customer churn. This model aims to forecast whether a customer will migrate to a different provider in a specific month, using past data. The ultimate goal is to prevent customer churn by accurately predicting such behavior.
Python packages used sklearn, Numpy, Pandas, Matplotlib and Seaborn
#SupervisedLearning, #PCA, #LogisticRegression, #DecisionTree, #RandomForest, #GradientDescent, #XGBoost, #LightGBM, #ClassImbalnce, #SMOTE
This case study focuses on the development and application of a Convolutional Neural Network (CNN) model for the early and accurate detection of melanoma, a type of skin cancer that is responsible for 75% of skin cancer-related deaths. The objective is to create a solution that can analyze images and alert dermatologists about the potential presence of melanoma, thereby potentially reducing the manual effort required in diagnosis.
Python packages used TensorFlow, Keras, Numpy, Pandas and Seaborn
#DeepLearning, #ANN, #CNN, #TensorFlow, #Keras, #CrossEntropy, #HyperparametersTunning #ImageRecognition
In this case study, we’re part of a tech team at a leading electronics company, known for its innovative smart TVs. We’re tasked with developing a unique feature that allows the TV to identify five specific user gestures, eliminating the need for a remote. The TV’s integrated webcam monitors these gestures, each triggering a distinct command:
These gestures are identified using various Deep Learning Models such as Conv3D, CNN combined with LSTM, CNN combined with GRU, and CNN combined with GRU (with trainable weights from Transfer Learning).
Python packages used TensorFlow, Keras, skimage, numpy
#DeepLearning, #CNN, #RNN, #LSTM, # #TensorFlow, #Keras, #mobilenet, #cv2, #ModelCheckpoint, #Conv3D, #Conv2D, #MaxPooling3D, #MaxPooling2D, #Dropout
The objective of this project is to develop a model that automatically classifies customer complaints based on the mentioned products and services, enabling quicker issue resolution. Using non-negative matrix factorization (NMF) for topic modeling, the project will identify patterns and recurring words in tickets, helping to determine key features for each category cluster and understand the topics of customer complaints.
Python packages used nltk, spacy, tqdm , pickle, wordcloud
#NLP, #NMF, #TfidfVectorizer #RegularExpressions #TopicModelling
The primary goal of this project is to create a fashion query response system that leverages Large Language Model (LLM) to deliver detailed, user-friendly answers to fashion-related questions. The system is designed to improve user experience by providing informative and contextually relevant responses, helping users find fashion items tailored to their preferences.
Python packages used OpenAI, chromadb, CrossEncoder, tiktoken, nltk, PIL
#GPT3.5, #OpenAI, #RAG, #Embeddings, #ChromaDB, #VectorStore, #HuggingFace, #GenAI #Flask
The objective of this project is to enable interactive Q&A with multiple PDF files using the Google Gemini Generative AI model. Users can upload PDFs, ask questions related to their content, and receive accurate responses based on the document information.
Python packages used Gemini, LangChain, chromadb, PyPDF2, streamlit, google-generativeai
#GeminiPro, #LangChain, #RAG, #Embeddings, #ChromaDB, #HuggingFace, #GenAI
The objective of this project is to develop a sentiment-based product recommendation system to enhance user experience and drive customer satisfaction. Users will benefit from personalized product suggestions based on their reviews and ratings.