My First Machine Learning Project
I found a fun dataset about the dogs of Zurich, Switzerland on Kaggle. I will build figures, classifiers, and other fun data related things around this dataset. I will be using Tensorflow, Numpy, and Pandas.
Pandas / matplotlib warmup
Located in stats.py, shows trending dogs and information about their owners
Linear classifier
Classifies the gender of the owner based on dog features.
accuracy: 0.686246
accuracy/baseline_label_mean: 0.313754
accuracy/threshold_0.500000_mean: 0.686246
auc: 0.5
global_step: 150
labels/actual_label_mean: 0.313754
labels/prediction_mean: 2.76343e-07
loss: 4.78862
precision/positive_threshold_0.500000_mean: 0.0
recall/positive_threshold_0.500000_mean: 0.0
Neural Network classifier
Work in progress
Dataset Notes
Since German is the official language of Zurich most of the columns are in German but the translations to English aren't too tricky
- ALTER -> Age
- GESCHLECHT -> Gender
- STADTKREIS -> City Quarter or District
- RASSE1 -> Dog's Primary Breed
- RASSE2 -> Dog's Secondary Breed
- GEBURTSJAHR_HUND -> Dog's Year of Birth
- GESCHLECHT_HUND -> Dog's Gender
- HUNDEFARBE -> Dog's Color