File size: 2,923 Bytes
32dea6f |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 |
# script to create a new (normalized) observation data point for the plane crash dataset
# author tbartolo
import joblib
import numpy as np
import gradio as gr
import os
resources_dir = "resources"
normalizers_dir = "normalizers"
algorithms_dir = "algorithms"
def master_fn(year, operator, location, ac_type, passengers, crew, algorithm):
def ac_type_normalizer(value):
hasher = joblib.load(os.path.join(resources_dir, normalizers_dir, "ac_type_hash_model.joblib"))
return hasher.transform([[value]]).toarray().flatten()
def location_normalizer(value):
hasher = joblib.load(os.path.join(resources_dir, normalizers_dir, "location_hash_model.joblib"))
return hasher.transform([[value]]).toarray().flatten()
def operator_normalizer(value):
hasher = joblib.load(os.path.join(resources_dir, normalizers_dir, "operator_hash_model.joblib"))
return hasher.transform([[value]]).toarray().flatten()
def year_normalizer(value):
kbins_discretizer = joblib.load(os.path.join(resources_dir, normalizers_dir, "year_qbins_model.joblib"))
return kbins_discretizer.transform([[value]]).flatten()
def passenger_crew_normalizer(value1, value2):
scaler = joblib.load(os.path.join(resources_dir, normalizers_dir, "passenger_crew_scaler_model.joblib"))
return scaler.transform([[value1, value2]])[0]
observation = np.concatenate((year_normalizer(year), operator_normalizer(operator), location_normalizer(location),
ac_type_normalizer(ac_type), passenger_crew_normalizer(passengers, crew)))
result = float
if algorithm == "SVR":
svr = joblib.load(os.path.join(resources_dir, algorithms_dir, "svr_model.joblib"))
result = svr.predict([observation])[0]
elif algorithm == "KNN":
knn = joblib.load(os.path.join(resources_dir, algorithms_dir, "knn_model.joblib"))
result = knn.predict([observation])[0]
elif algorithm == "Isolation Forest":
random_forest = joblib.load(os.path.join(resources_dir, algorithms_dir, "random_forest_model.joblib"))
result = random_forest.predict([observation])[0]
else:
raise ValueError('Unknown algorithm: {}'.format(algorithm))
return round(result, 2)
demo = gr.Interface(
master_fn,
[
"number",
"text",
"text",
"text",
"number",
"number",
gr.Radio(["KNN", "Isolation Forest", "SVR"])
],
"number",
title="Flight Crash Fatality Prediction",
description="Enter the flight details and the algorithm to perform the prediction. Note that the purpose of this "
"effort is strictly to showcase the strengths and weaknesses amongst the 3 techniques used, "
"and by no means claims any sort of accuracy in its predictions.",
)
if __name__ == "__main__":
demo.launch(show_api=False)
|