Discover @ner-aim Open Source projects

Pottapatri (@ner-aim)

ner-aim

Stars
43
Global Rank 379,383 (Top 14 %)
Followers 15
Following 39
Registered over 6 years ago
Most used languages

Jupyter Notebook
43.8 %

C#
25.0 %

CSS
6.3 %

Shell
6.3 %

Python
6.3 %

JavaScript
6.3 %

ASP.NET 6.3 %
Location 🇮🇳 India
Country Total Rank 13,221
Country Ranking

ASP.NET 47

C#
596

Jupyter Notebook
4,435

Shell
4,868

CSS
8,401

TwinMemories-Prototype

ner-aim.github.io

Visit my portfolio: https://ner-aim.github.io

HR-Analytics-Dashboard

Facial-Expression-Detection

Jupyter Notebook

A-Tile-Game

A platformer game.

Super-Sparty-Bros

This game is the part of a course and a clone to Super Mario Bros.

Reinforcement-Learning-using-Unity

Reinforcement learning with Unity

Capital-One-Cards-Fraud-Detection-MLOps

Jupyter Notebook

Course-Projects

academic-kickstart

Retreving-Financial-Estimates-using-Estimize

This project revolves around getting the financial estimates for a few companies of the financial year 2019

Jupyter Notebook

Capital-One-Credit-Card-Fraud-Predicition

Jupyter Notebook

Color-Game

A basic color game where you have to guess the color of the given RGB composition.

TicTacToe

Just a simple TicTacToe game made using python tkinter

Credit-Card-Approval-Prediction

Jupyter Notebook

US-Prevailing-Wages

Operated with a team of 5 and preprocessed data using Pandas by clearing problematic rows and brought consistency and leveraged 10 common features Generated visualizations using Plotly for gathering hidden insights of wage growth, geographic, NAICS, job experience

Jupyter Notebook

Predicting-seasurface-temperatures

Comparing all 3 models, using the Residual Mean Squared Error as our metric, we see that the BSTS model performs the best with the lowest RMSE score. We suspect that the Bayesian approach worked so well because it can infer causal impacts within time series. In this case, whether 1 or more layers greatly influenced the temperatures of the layers beneath it. In our coefficient analysis for a model size of 1 we confirmed this was the case. We faced a few issues in this project. For the SARIMA models, we started out with a trial and error method to determine the order of the seasonal model required to fit the data. This method proved difficult since there are numerous combinations of numbers that could have made the model. Reading the ACF and PACF model and using that as a guide resulted in producing the best model we have provided above. Another issue we ran into was trying to configure the predict method to work with our data. The predict method produced predictions for the entire dataset instead of just future data. Using the one step ahead prediction and the forecast function was able to rectify this error.

Jupyter Notebook