shubhangperiwal Goto Github PK

active_ingredient_measure

This assignment seeks to use a regularized logistic regression model with L1 penalty function to predict the amount of active ingredient given the input NIR data. We then need to train and validate the model using data x and y and then compute the performance of the model x_test with y_test where x, x_test are predictor variables and y, y_test are response variables. We then need to find the misclassification rate of this model.

aiproject

bidirectional with ui london underground

algorithms

bla

authoridentification

To identify author name from a few words of text using multiple machine learning, NLP and neural network models

bayesiananalysis_misprint

Misprint Analysis using bayesian inference

bayesian_assignment4

bloodbank

Website for a bloodbank with backend system design

c-c_codes

It includes a few algorithms that i have coded in c/c++, it also includes some work on opencl an mpi

digit_classification1

The file data_usps_digits.RData contains data recording handwritten digits from the United States Postal Service (USPS). More in details, the data consists of grayscale (16x16) grid representations of image scans of the digits “0” through “9” (10 digits). We need to use a multilayer neural network with 2 hidden layers to predict the type of digit. We then need to perform analysis of the output. This would be followed by an analysis for 3 hidden layers.

digit_classification2

Identifying the best neural network for digit identification using sampling of multiple neural network architectures

distancecalculator

mfp

mathematica_genre_classification

Music genre classification using multiple machine learning algorithms and their comparison(Mathematica, Wolfram Alpha)

mathematica_practice

music_clustering

This assignment seeks to us to do a complete cluster analysis of the Spotify audio features data using k-means followed by k-medoids. The dataset data_spotify_songs.rda contains data about audio features for a collection of songs. The songs belong to three genres: acoustic, pop and rock. The dataset contains 239 songs. The properties that we can use for clustering are song duration, danceability, energy, liveness, speech, tempo and audio valence.

price_analysis

Price analysis using multiple features (regression and pre-processing) followed by model performance analysis

r_practice

R programming

satellite_image_classification

Satellite image classification using logistic regression and random forest to predict the classification of images. Followed by their comparison and performance of the best model on the data.

solar_system_count

Comparison of multiple supervised machine learning algorithms such as Random forest, linear regression, multinomial regression, classification trees, bagging, boosting, SVM (state vector machine), one of them was also using polling among three algorithms. This is preceded by dimension reduction technique such as PCA (reducing 57 dimensions to 31), this helps in a faster training, testing and better results in a few cases. Furthermore, I have compared all algorithms mentioned approximately 50 times using random data sampling. As well as done a polling between top 3 algorithms to show whether they are able to perform better when they’re used together.

sorting-algorithm-visualizer

spam_detection

Spam detection using regression followed by detailed analysis

sports_activity_recognition

identification of sports data using multiple neural network architectures and CNN