karthikreddyparla Goto Github PK

analising_toyota_camry_reviewsnatural_language_processing

Scraped,cleaned and assimilated Toyota Camry car reviews as input to star rating prediction model using Natural language processing. Derived sentiment score(afinn lexicon ) and tags from the reviews to classify and quantify the reviews.

bitcoin_prices_prediction

Predicted the future of bitcoin prices by applying cross validation and grid search to derive the most efficient model. (ridge, lasso, KNN, linear ,polynomial regression and SVM)

deep_learning_to_predict_houseload

Studied the trend of overall load and divided it into three classes, low, high and average efficient labels. Then trained a deep learning model using KerasClassifier and three-layered baseline model with relu and softmax as activation functions to predict the label in Python Jupiter book Built explanatory deep learning model to predict the house heating and cooling load using tensor-flow and keras with 93% accuracy

dimension_reduction_effect_on_rsquare

Used PCA and manifold to reduce the dimensions from 257 to 83 on cardiac_arrhythmia data set. Compared Logistic Regression, KNN, and linear SVM models on both 257 and 83 components and found that there is not much difference in the R2. This is because of the co- relation among the components.

sentiment-analysis-on-toyota-camry-comments

snakegame

social_network_analysis

Social Network Analysis: Studied behaviour of a email network by finding the one and two hop neighbours, betweenness, indegree, centrality of a node.

test

xor_gate_using_neuralnetwork-

Wrote a function to depict the input to a neuron and included sigmoid activation function to get required output.Constructed a three-layer neural network to function as XOR gate using the neurons above and included a step activation function to the final neuron to get the output in 0 and 1.