MohanBisunke
Machine Learning, Deep Learning , Natural Language Processing , Large Language Models, Time Series on Python.
Nepal
Pinned Repositories
analmoly-detection-in-timeseries-with-IsolationForest-and-AutoEncoders
With the use of autoencoders in ECG dataset, the reconstructed image (by decoder side) of ECG is used to see if there is anamoly or not. With IsolationForest, a traffic based nyc taxi is evaluated in diffrent timeframe to see if there is fluctuating demands
Appliances-Energy-Usuage-VAR-TimeSeries
The VAR model is used to forecast the appliances energy on the previous usage history. The data were first tested using adfuller test, granger casuality test. The lag value of 7 was determined for VAR model after running it iteratively for values upto 48.
book-recommeder
popularity based and collaborative filtered based recommendation system
Email_Multiclass_classification_LSTM
The built model(LSTM) is able to Categorize emails into Primary, Spam, Promotions and Social. Because of dataset taken from individual, it's not uploaded for privacy concern.
face-mask-detection-with-VGG16-CNN-and-Haarcascade-openCV-
Haar Cascade frontalface is used for detecting facial region from frame and sent to VGG16 model (tuned for binary classification) to detect mask on face.
Haarcascade-based-face-detection-using-KNN-in-OpenCV
This programs capture real-time video frames, utilizing Haar cascades for precise facial detection and K-Nearest Neighbors for identification of people from trained dataset. Folder face_dataset is not added intentionally.
motion-detection-of-people-with-contour-plot-in-opencv
The program reads frames from a video file, compares consecutive frames to identify areas of motion, and highlights them with bounding boxes. When motion is detected, it displays the processed video feed in a window, and simultaneously saves the video with the detected motion to an output file.
movie-recommedation
Movie Recommender based on CountVectorizer and Cosine Similarity. Given one movies, similar five movies will be recommended
NYC-Energy-Consumption-TimeSeries-LSTM
NYC Energy consumption is time series based problem, model is trained on Deep learning LSTM networks to predict the usuage and demand of energy.
NYC_Enenrgy_Consumption_with_AutoARIMA
This is simple workflow for predicting the future usuage of energy demand based on time series analysis with Auto ARIMA.
MohanBisunke's Repositories
MohanBisunke/analmoly-detection-in-timeseries-with-IsolationForest-and-AutoEncoders
With the use of autoencoders in ECG dataset, the reconstructed image (by decoder side) of ECG is used to see if there is anamoly or not. With IsolationForest, a traffic based nyc taxi is evaluated in diffrent timeframe to see if there is fluctuating demands
MohanBisunke/Appliances-Energy-Usuage-VAR-TimeSeries
The VAR model is used to forecast the appliances energy on the previous usage history. The data were first tested using adfuller test, granger casuality test. The lag value of 7 was determined for VAR model after running it iteratively for values upto 48.
MohanBisunke/book-recommeder
popularity based and collaborative filtered based recommendation system
MohanBisunke/Email_Multiclass_classification_LSTM
The built model(LSTM) is able to Categorize emails into Primary, Spam, Promotions and Social. Because of dataset taken from individual, it's not uploaded for privacy concern.
MohanBisunke/face-mask-detection-with-VGG16-CNN-and-Haarcascade-openCV-
Haar Cascade frontalface is used for detecting facial region from frame and sent to VGG16 model (tuned for binary classification) to detect mask on face.
MohanBisunke/Haarcascade-based-face-detection-using-KNN-in-OpenCV
This programs capture real-time video frames, utilizing Haar cascades for precise facial detection and K-Nearest Neighbors for identification of people from trained dataset. Folder face_dataset is not added intentionally.
MohanBisunke/motion-detection-of-people-with-contour-plot-in-opencv
The program reads frames from a video file, compares consecutive frames to identify areas of motion, and highlights them with bounding boxes. When motion is detected, it displays the processed video feed in a window, and simultaneously saves the video with the detected motion to an output file.
MohanBisunke/movie-recommedation
Movie Recommender based on CountVectorizer and Cosine Similarity. Given one movies, similar five movies will be recommended
MohanBisunke/NYC-Energy-Consumption-TimeSeries-LSTM
NYC Energy consumption is time series based problem, model is trained on Deep learning LSTM networks to predict the usuage and demand of energy.
MohanBisunke/NYC_Enenrgy_Consumption_with_AutoARIMA
This is simple workflow for predicting the future usuage of energy demand based on time series analysis with Auto ARIMA.
MohanBisunke/Opencv-Core-operation-and-image-processing
Basic operations in opencv with image processing like read, write ,display, thresholding, smoothing, transformation, edge detection, contours ,histograms , hough line , watershed etc
MohanBisunke/quora_duplicate_question_pair
The built model is for checking whether the questions are duplicate or not given two questions.
MohanBisunke/romanized-nepali-text-sentiment-analysis
The built model (Naive Bayes: Multinomial, Gaussian and Bernoulli combined voted) is Sentiment Analyzer for Romanized Nepali Text (Nepali words spoken in English Letters).The model predict either positive or negative
MohanBisunke/Time-Series-basic-along-with-Scipy
This includes scipy basics( stats,integrate , linear algebra, minimization , interpolation, basics ), time series basic ideas like Handling missing values, stationary test, moving average , decomposition, data preparation for LSTM , sequence models etc.