jones5am

Connecting_IoT_to_AutoML

One of the most common complaints for Data Science & AI in general is that time to implementation for any practical business application is long and arduous. Among others, this is often due to reasons such as the lack of valuable & correctly pre-processed data for modeling, the need for connectivity between the source of the data and the predictive model, and the lengthy steps of model selection, through hyperparameter tuning & validation. This is no different in the context of manufacturing, where a diverse array of machines, widgets, locations & often low fault tolerances result in the need for customized models that cannot be cost effectively developed through traditional means. This project’s goal will therefore be showcasing how low cost IoT devices & sensors can be combined with cloud computing storage and AutoML packages to very quickly produce powerful machine learning algorithms for manufacturing at scale. The goal of this project is not however, to produce the best classifier model possible on the manufacturing dataset which is used to showcase the final AutoML portion.

DataMining_Commodities

repository for data mining project on commodities prices

DataMining_WorldMilitarySpending

repository for data mining project on world military spending

f18_ds_nlp

Malware_Classification_Using_Deep_Learning

Data Science Capstone project on creating a deep learning algorithm that can classify compressed images of known malware binary streams

Network_DOS_Attack_Detection

Using the 1998 DARPA Intrusion Detection Evaluation dataset I configured a Random Forest model for anomaly detection

online-cv

A minimal Jekyll Theme to host your resume (CV)

PredictiveModel_SeattleHousingPrices

Predictive Model on Seattle Housing Prices using regsubsets method for feature selection

PUBG_Data

repository for my randomly sampled PUBG data to be used in OutSchool Class

Regsubset_FeatureSelection_HealthData

Regsubsets method for feature selection on Health Data

jones5am/Network_DOS_Attack_Detection

Using the 1998 DARPA Intrusion Detection Evaluation dataset I configured a Random Forest model for anomaly detection

jones5am/DataMining_WorldMilitarySpending

repository for data mining project on world military spending

jones5am/Malware_Classification_Using_Deep_Learning

Data Science Capstone project on creating a deep learning algorithm that can classify compressed images of known malware binary streams

jones5am/Connecting_IoT_to_AutoML

One of the most common complaints for Data Science & AI in general is that time to implementation for any practical business application is long and arduous. Among others, this is often due to reasons such as the lack of valuable & correctly pre-processed data for modeling, the need for connectivity between the source of the data and the predictive model, and the lengthy steps of model selection, through hyperparameter tuning & validation. This is no different in the context of manufacturing, where a diverse array of machines, widgets, locations & often low fault tolerances result in the need for customized models that cannot be cost effectively developed through traditional means. This project’s goal will therefore be showcasing how low cost IoT devices & sensors can be combined with cloud computing storage and AutoML packages to very quickly produce powerful machine learning algorithms for manufacturing at scale. The goal of this project is not however, to produce the best classifier model possible on the manufacturing dataset which is used to showcase the final AutoML portion.

jones5am/DataMining_Commodities

repository for data mining project on commodities prices

jones5am/f18_ds_nlp

jones5am/online-cv

A minimal Jekyll Theme to host your resume (CV)

jones5am/PredictiveModel_SeattleHousingPrices

Predictive Model on Seattle Housing Prices using regsubsets method for feature selection

jones5am/PUBG_Data

repository for my randomly sampled PUBG data to be used in OutSchool Class

jones5am/Regsubset_FeatureSelection_HealthData

Regsubsets method for feature selection on Health Data

jones5am/TemporalAnalysis_ChicagoCrime-

Temporal Analysis on Chicago Crime data