gravatar

harpomaxx

Carlos Catania (AKA Harpo)

Recently Published

Elasticnet baseline for FeNi
Random forest basline for FeNi
Ranger Time-Series
Random Forests for botnet detection using graph features
Random Forests Confidence Intervals
AI-1 Intro dplyr
Opinions about IA. AWordcloud example
EDA Vines 2022 (5 features)
EDA Vines 2022 (20 features)
reflex vs random agent
Dropout efect in validation
Labeling Strategies
DBplyr examples
SHAP Value Summary Plot for Goat Behavior
Absolute Average SHAP value contribution to ethogram considering 4 activities: (G)razing, (R)esting, (W)alking and (G)razing (M)imosa.
caret for random forest and knn example
Ejemplo Histogramas
Reticulate R & Python
prophet
Calculando Media y Desviación de algoritmos de busquedas no informada
MeLI 2021
Keras learning curves for sequence classification
Sequence classification models
data-centric vs data-model approach for sequence classification
EDA vines
Correlation Matrix Vines
CTU19 Sampling size analysis and creation
CTU19 PCA representation (Sequences)
IZ ICL POSITION UNCERTAINTY ANALYSIS
IOL RETRO POSITION CART MULTIPLE FEATURES
IOL RETRO POSITION CART RETESTING
IOL RETRO POSITION CART
Ganado Caprino bis 2020
catboost simple example
catboost vs random forest for goat behavior classification
catboost evaluation on goat behavior classification problem
IZL ICL POSITION ANALYSIS
ICL Postop EDA
Document
CTU13 vs CTU19 UMAP representation
Yet Another COVID19 Plot. Deads per 100k population
DECONV Challenge 20000 features RF fine grain
DECONV Challenge 20000 features RF 5 (newmix)
DECONV Challenge 20000 features RF 4
CTU19- LSTM results
CTU19- LSTM results (PCA)
DECONV Challenge 20000 features RF 3
DECONV Challenge 20000 features RF 2
DevCon Challenge 2020
DEVCON Challenge
Yet Another COVID19 Projection (2)
Peng et al 2020 SEIQRDP compartment model projections for several countries.
COVID19: Number of Deaths since first death was reported. (Other countries)
COVID19: deaths per country every 100k population (since first death)
COVID19: deaths per country every 100k population
Yet Another COVID19 Projection
Peng et al. 2020, SEIQRDP compartment model projections for several countries.
caret machine learning y el titanic
COVID19 Forecast for ARGENTINA
The models were fitted with observed data 15 days from today and projected for the next 7 days
ctu13bis dataset
ASAI 2020 results
ASAI 2020 results
Evaluation of MC-CNN on Full domains
Analysis of MC-NN for DGA detection
wine dataset meetup Data Science
MC-DGA analysis
Fenoles
big CNN tuning results
Descriptive Analysis of CTU Dataset
Original NN vs cacic_python NN
WB NN model comparison heatmap
Analisis Dataset Ganado Caprino
CNN base evaluated on normal domains
JISA 2018
Rainfall prediction for Mendoza - daily
AGRANDA 2018
Analysis of CNN results for conficker malware
LCC analysis
DAAC premilinar analysis
camera identification
Spoiler data analysis
Testing Keras for R
I've just followed the instruction from http:///keras.rstudio.com. Everything worked!
WB profile classifier
WB profile classifer
ejercicios Complejidad 2017
ALGO I Trabajo Practico Nro I
wp pricing
clustering trivial example
learning rate by label
Normal-TCP-HTTP
Botnet-TCP-SMTP-Attempt-SPAM distribution
Exploring the CTU 13 labeled Botnet Dataset (new)
HACKATON AM 2017 Clases desbalanceadas
Como lidiar con el problema de clases desbalanceadas usando caret
HACKATON AM 2017 Exploracion Inicial
Una exploracion inicial del conjunto de datos del arbolado publico utilizando herramientas comunes de R.
HACKATON AM 2017 Ejemplo de envio
Un sencillo ejemplo de como generar un archivo para enviar al desafio de Kaggle de la Hackaton 2017 de Aprendizaje de Maquinas FRM - UTN - Mendoza - Argentina
AM 2017 HACKATON
Descripción del desafio de Kaggle para la materia Aprendizaje Automatico 2017 de la FRM UTN - Mendoza - Argentina
Arbolado Publico Mendoza
WB 1h Client Profile Generation (last 8 days)
WB 24h Client Profile Generation (last 21 days)
ID3 basic operations
Simple Bag of Word
CTU-13 Labeling
AM2017 L0 (tidyverse)
Laboratorio 0 con algunos toques del tidyverse
WB DGA detection
Exploring the CTU 13 labeled Botnet Dataset
NN results for word-based DGA
wb DGA
Character Relative Frequency Distribution (percent)
DGA vs Legit Proportion on 7hs of WB data
NN DGA results
Results on 400000 domains (50% DGA, 50% legit)
DGA dataset
Customer Segmentation II
EAAI 2016 Brief v2
Customer Segmentation
Preliminary Client Analysis
Action Plot
rankID action plot
rankID action plot
GISTAM 2017
EAAI 2016 Brief
Fitting Cmorph to Station data
Cmorph Precipitation days for MZA
Connection Heatmap
Testing a simple RF model for helping labeling datasets
The idea is using the information provided by labeled data to construct a classifier that gives the probability of being malicious traffic.
EAAI 2016 Report 4
Heatmap evaluation on full dataset
heatmap evaluation
Station vs CMORPH
Evaluation of Jay Jacobs DGA features on Andrew Datasets
https://github.com/jayjacobs/dga/
MA vs TDMLR on xbow
MA vs TDMLR on gladiator
EAAI 2016 report 2-xbow
EAAI 2016 report 2
Cmorph Wavelet Analisys for CABA
Cmorph Wavelet Analisys
Trajectories for Melbourne Dataset
Basic Reshaping (II)
Dynamic B
The ECML-PKDD Network Classification Challenge correlation matrix heatmap
Correlation matrix heatmap for precipitation prediction
Technical Report 1
Results after evaluation Gladiator IMU using several IMUs as target
AccX sensor vs TD-MLR
Melbourne IMU dataset
Melbourne Dataset Sensor AccX
Basic Melting
Basic Melting
size-10 patterns distribution plot
size-10 patterns present in at most 50 different connections, ordered from largest to smallest