Iris¶

This example uses the iris dataset to predict the species of various flowers based on their physical features.

PetalLengthCm: Petal Length in cm
PetalWidthCm: Petal Width in cm
SepalLengthCm: Sepal Length in cm
SepalWidthCm: Sepal Width in cm
Species: The Flower Species (Setosa, Virginica, Versicolor)

We will follow the data science cycle (Data Exploration - Data Preparation - Data Modeling - Model Evaluation - Model Deployment) to solve this problem.

Initialization¶

This example uses the following version of vastorbit:

import vastorbit as vo

vo.__version__

Connect to VAST. This example uses an existing connection called VASTDSN . For details on how to create a connection, see the Connection tutorial.

You can skip the below cell if you already have an established connection.

vo.connect("VASTDSN")

Let’s create a VastFrame of the dataset.

from vastorbit.datasets import load_iris

iris = load_iris()
iris.head(5)

	123 sepallengthcm Decimal(5,2) 100%	123 sepalwidthcm Decimal(5,2) 100%	123 petallengthcm Decimal(5,2) 100%	123 petalwidthcm Decimal(5,2) 100%	Abc species Varchar(30) 100%
1	5.1	3.5	1.4	0.2	Iris-setosa
2	4.9	3.0	1.4	0.2	Iris-setosa
3	4.7	3.2	1.3	0.2	Iris-setosa
4	4.6	3.1	1.5	0.2	Iris-setosa
5	5.0	3.6	1.4	0.2	Iris-setosa

Data Exploration and Preparation¶

Let’s explore the data by displaying descriptive statistics of all the columns.

iris.describe(method = "categorical", unique = True)

	dtype	count	top	top_percent	unique
"sepallengthcm"	decimal(5,2)	150	5.0	6.667	35.0
"sepalwidthcm"	decimal(5,2)	150	3.0	17.333	23.0
"petallengthcm"	decimal(5,2)	150	1.5	9.333	43.0
"petalwidthcm"	decimal(5,2)	150	0.2	18.667	22.0
"species"	varchar(30)	150	Iris-virginica	33.333	3.0

We don’t have much data here, but that’s okay; since different flower species have different proportions and ratios between those proportions, we can start by making ratios between each feature.

We’ll need to use the One-Hot Encoder on the Species to get information about each species.

iris["Species"].one_hot_encode(drop_first = False)
iris["ratio_pwl"] = iris["PetalWidthCm"] / iris["PetalLengthCm"]
iris["ratio_swl"] = iris["SepalWidthCm"] / iris["SepalLengthCm"]

We can draw the correlation matrix (Pearson correlation coefficient) of the new features to see if there are some linear links.

iris.corr()

The Iris setosa is highly linearly correlated with the petal length and the sepal ratio. We can see a perfect separation using the two features (though we can also see this separation the petal length alone).

iris.scatter(
    columns = ["PetalLengthCm", "ratio_swl"],
    by = "Species",
)

We can we a clear linear separation between the Iris setosa and the other species, but we’ll need more features to identify the differences between Iris virginica and Iris versicolor.

iris.scatter(
    columns = [
        "PetalLengthCm",
        "PetalWidthCm",
        "SepalLengthCm",
    ],
    by = "Species",
)

Our strategy is simple: we’ll use two Linear Support Vector Classification (SVC): one to classify the Iris setosa and another to classify the Iris versicolor.

Machine Learning¶

Let’s build the first LinearSVC to predict if a flower is an Iris setosa.

from vastorbit.machine_learning.vast import LinearSVC
from vastorbit.machine_learning.model_selection import cross_validate

predictors = ["PetalLengthCm", "ratio_swl"]
response = "Species_Iris-setosa"
model = LinearSVC("svc_setosa_iris", max_iter = 1000)
cross_validate(model, iris, predictors, response)

	auc	prc_auc	accuracy	log_loss	precision	recall	f1_score	mcc	informedness	markedness	csi	time
1-fold	1.0	0.17571428571428582	1.0	0.1948176137681635	1.0	1.0	1.0	1.0	1.0	1.0	1.0	0.39058709144592285
2-fold	1.0	0.17387820512820507	1.0	0.17402997622758762	1.0	1.0	1.0	1.0	1.0	1.0	1.0	0.5655460357666016
3-fold	1.0	0.1881127450980392	1.0	0.19978879457548462	1.0	1.0	1.0	1.0	1.0	1.0	1.0	0.3199501037597656
avg	1.0	0.17923507864684338	1.0	0.18954546152374527	1.0	1.0	1.0	1.0	1.0	1.0	1.0	0.42536107699076336
std	0.0	0.006322052366843215	0.0	0.01115723599691346	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.10323521042609826

Our model is excellent. Let’s build it using the entire dataset.

model.fit(iris, predictors, response)

Let’s plot the model to see the perfect separation.

model.plot()

We can add this probability to the VastFrame.

model.predict_proba(iris, name = "setosa", pos_label = 1)

	123 sepallengthcm Decimal(5,2) 100%	123 sepalwidthcm Decimal(5,2) 100%	123 petallengthcm Decimal(5,2) 100%	123 petalwidthcm Decimal(5,2) 100%	Abc species Varchar(30) 100%	123 species_Iris-setosa Bool 100%	123 ratio_pwl Decimal(13,8) 100%	123 ratio_swl Decimal(13,8) 100%	123 setosa Double 100%
1	5.1	3.5	1.4	0.2	Iris-setosa	1	0.14285714	0.68627451	0.7533862310328937
2	4.9	3.0	1.4	0.2	Iris-setosa	1	0.14285714	0.6122449	0.7347100983001176
3	4.7	3.2	1.3	0.2	Iris-setosa	1	0.15384615	0.68085106	0.7673806330944822
4	4.6	3.1	1.5	0.2	Iris-setosa	1	0.13333333	0.67391304	0.7342632994790236
5	5.0	3.6	1.4	0.2	Iris-setosa	1	0.14285714	0.72	0.761595468695358
6	5.4	3.9	1.7	0.4	Iris-setosa	1	0.23529412	0.72222222	0.7134820101455942
7	4.6	3.4	1.4	0.3	Iris-setosa	1	0.21428571	0.73913043	0.7661678386279932
8	5.0	3.4	1.5	0.2	Iris-setosa	1	0.13333333	0.68	0.7358342158526154
9	4.4	2.9	1.4	0.2	Iris-setosa	1	0.14285714	0.65909091	0.7466324281606391
10	4.9	3.1	1.5	0.1	Iris-setosa	1	0.06666667	0.63265306	0.7234590908839088
11	5.4	3.7	1.5	0.2	Iris-setosa	1	0.13333333	0.68518519	0.737167699846424
12	4.8	3.4	1.6	0.2	Iris-setosa	1	0.125	0.70833333	0.7267037644243937
13	4.8	3.0	1.4	0.1	Iris-setosa	1	0.07142857	0.625	0.7379915215420295
14	4.3	3.0	1.1	0.1	Iris-setosa	1	0.09090909	0.69767442	0.7996161340009732
15	5.8	4.0	1.2	0.2	Iris-setosa	1	0.16666667	0.68965517	0.7840216084848081
16	5.7	4.4	1.5	0.4	Iris-setosa	1	0.26666667	0.77192982	0.7588254032478668
17	5.4	3.9	1.3	0.4	Iris-setosa	1	0.30769231	0.72222222	0.7770233333481225
18	5.1	3.5	1.4	0.3	Iris-setosa	1	0.21428571	0.68627451	0.7533862310328937
19	5.7	3.8	1.7	0.3	Iris-setosa	1	0.17647059	0.66666667	0.6981988028179743
20	5.1	3.8	1.5	0.3	Iris-setosa	1	0.2	0.74509804	0.7522584560908878

Let’s create a model to classify the Iris virginica.

predictors = [
    "PetalLengthCm",
    "SepalLengthCm",
    "SepalWidthCm",
    "PetalWidthCm",
    "ratio_pwl",
    "ratio_swl",
]
response = "Species_Iris-virginica"
model = LinearSVC("svc_virginica_iris", max_iter = 1000)
cross_validate(model, iris, predictors, response)

	auc	prc_auc	accuracy	log_loss	precision	recall	f1_score	mcc	informedness	markedness	csi	time
1-fold	1.0	0.24272727272727268	1.0	0.20065743001538366	1.0	1.0	1.0	1.0	1.0	1.0	1.0	0.3792917728424072
2-fold	1.0	0.2242324561403508	0.9583333333333334	0.239925695296195	1.0	0.8947368421052632	0.9444444444444444	0.914883536058052	0.8947368421052633	0.935483870967742	0.8947368421052632	0.14366984367370605
3-fold	0.9966329966329966	0.21043076099835645	0.9411764705882353	0.23769605483224418	0.9411764705882353	0.8888888888888888	0.9142857142857143	0.8703882797784892	0.8585858585858586	0.8823529411764706	0.8421052631578947	0.30555009841918945
avg	0.9988776655443322	0.22579682995532666	0.9665032679738562	0.22609306004794094	0.9803921568627452	0.9278752436647174	0.9529100529100529	0.9284239386121804	0.9177742335637072	0.9392789373814042	0.912280701754386	0.2761705716451009
std	0.0015872206087240053	0.013231316803673081	0.024699700094495056	0.018008725302187445	0.027729677693590107	0.051055753815628845	0.03550102898881748	0.05377301913717624	0.059986281219858315	0.04810411977692991	0.06564311204866566	0.09840999549977007

We have another excellent model. Let’s add it to the VastFrame.

model.fit(iris, predictors, response)
model.predict_proba(iris, name = "virginica", pos_label = 1)

	123 sepallengthcm Decimal(5,2) 100%	123 sepalwidthcm Decimal(5,2) 100%	123 petallengthcm Decimal(5,2) 100%	123 petalwidthcm Decimal(5,2) 100%	Abc species Varchar(30) 100%	123 species_Iris-setosa Bool 100%	123 ratio_pwl Decimal(13,8) 100%	123 ratio_swl Decimal(13,8) 100%	123 setosa Double 100%	123 virginica Double 100%
1	5.1	3.5	1.4	0.2	Iris-setosa	1	0.14285714	0.68627451	0.7533862310328937	0.0007296201703394868
2	4.9	3.0	1.4	0.2	Iris-setosa	1	0.14285714	0.6122449	0.7347100983001176	0.0014058511173605517
3	4.7	3.2	1.3	0.2	Iris-setosa	1	0.15384615	0.68085106	0.7673806330944822	0.0011623041879743321
4	4.6	3.1	1.5	0.2	Iris-setosa	1	0.13333333	0.67391304	0.7342632994790236	0.0018303349093362288
5	5.0	3.6	1.4	0.2	Iris-setosa	1	0.14285714	0.72	0.761595468695358	0.0007114409374903144
6	5.4	3.9	1.7	0.4	Iris-setosa	1	0.23529412	0.72222222	0.7134820101455942	0.0008412646621078331
7	4.6	3.4	1.4	0.3	Iris-setosa	1	0.21428571	0.73913043	0.7661678386279932	0.001396973628006954
8	5.0	3.4	1.5	0.2	Iris-setosa	1	0.13333333	0.68	0.7358342158526154	0.0009996800043388435
9	4.4	2.9	1.4	0.2	Iris-setosa	1	0.14285714	0.65909091	0.7466324281606391	0.002270472963791192
10	4.9	3.1	1.5	0.1	Iris-setosa	1	0.06666667	0.63265306	0.7234590908839088	0.0012246022779389021
11	5.4	3.7	1.5	0.2	Iris-setosa	1	0.13333333	0.68518519	0.737167699846424	0.0005461931729756654
12	4.8	3.4	1.6	0.2	Iris-setosa	1	0.125	0.70833333	0.7267037644243937	0.0013341977091354206
13	4.8	3.0	1.4	0.1	Iris-setosa	1	0.07142857	0.625	0.7379915215420295	0.0012741245386073093
14	4.3	3.0	1.1	0.1	Iris-setosa	1	0.09090909	0.69767442	0.7996161340009732	0.0012243303583175569
15	5.8	4.0	1.2	0.2	Iris-setosa	1	0.16666667	0.68965517	0.7840216084848081	0.00019672232223855336
16	5.7	4.4	1.5	0.4	Iris-setosa	1	0.26666667	0.77192982	0.7588254032478668	0.0003086020837643312
17	5.4	3.9	1.3	0.4	Iris-setosa	1	0.30769231	0.72222222	0.7770233333481225	0.00047993869640503033
18	5.1	3.5	1.4	0.3	Iris-setosa	1	0.21428571	0.68627451	0.7533862310328937	0.000868941342595363
19	5.7	3.8	1.7	0.3	Iris-setosa	1	0.17647059	0.66666667	0.6981988028179743	0.0006185188604478436
20	5.1	3.8	1.5	0.3	Iris-setosa	1	0.2	0.74509804	0.7522584560908878	0.000739172518255419

Let’s evaluate our final model (the combination of two LinearSVC).

iris.case_when(
    "prediction",
    iris["setosa"] > 0.5, "Iris-setosa",
    iris["virginica"] > 0.5, "Iris-virginica",
    "Iris-versicolor",
)
iris["score"] = (iris["Species"] == iris["prediction"])

iris["score"].avg()

We have a great model with an accuracy of 96% on an entirely balanced dataset.

Conclusion¶

We’ve solved our problem in a pandas-like way, all without ever loading data into memory!