Dimentionality reduction

This project addresses the challenge of high-dimensional data in image classification. It explores various classification models and dimensionality reduction techniques to achieve an accurate and efficient solution. The goal is to develop a robust binary classification model and demonstrate its performance on new data. The dataset consists of 28x28 grayscale images from the Fashion MNIST dataset.

Approach

The project follows these steps:

Load and split the data into training, validation, and test sets.
Perform exploratory data analysis (EDA) to understand the dataset characteristics.
Apply Support Vector Machine (SVM), Naive Bayes, and Linear Discriminant Analysis (LDA) models.
- Discuss the suitability of each model for this task.
- Tune key hyperparameters.
- Experiment with data standardization and normalization.
- For SVM, test at least two different kernel functions.
- Analyze and comment on the results for each model.
Apply Principal Component Analysis (PCA) and Locally Linear Embedding (LLE) for dimensionality reduction.
- Re-evaluate the models with reduced dimensions to improve performance.
- Determine the optimal number of components for each reduction method.
- Analyze and comment on the results.
Select the best-performing model and estimate its accuracy on unseen data.

import pickle
from itertools import chain
from pathlib import Path

import matplotlib.pyplot as plt
import numpy as np
import pandas as pd
from sklearn.base import BaseEstimator
from sklearn.base import ClassifierMixin
from sklearn.decomposition import PCA
from sklearn.discriminant_analysis import LinearDiscriminantAnalysis
from sklearn.manifold import LocallyLinearEmbedding
from sklearn.metrics import ConfusionMatrixDisplay
from sklearn.metrics import PrecisionRecallDisplay
from sklearn.metrics import RocCurveDisplay
from sklearn.metrics import accuracy_score
from sklearn.metrics import classification_report
from sklearn.model_selection import ParameterGrid
from sklearn.model_selection import train_test_split
from sklearn.naive_bayes import GaussianNB
from sklearn.preprocessing import FunctionTransformer
from sklearn.preprocessing import Normalizer
from sklearn.preprocessing import StandardScaler
from sklearn.svm import SVC
from sklearn.svm import LinearSVC
from sklearn.utils.multiclass import unique_labels
from sklearn.utils.validation import check_array
from sklearn.utils.validation import check_is_fitted
from sklearn.utils.validation import check_X_y

RANDOM_STATE = 42
np.random.seed(RANDOM_STATE)
pd.set_option("display.precision", 4)
pd.set_option("display.max_columns", 10)
pd.set_option("display.max_rows", 100)
plt.style.use("default")
plt.rcParams["image.cmap"] = "Blues"
models = Path("models")

Data

train_file = "train.csv"
eval_file = "evaluate.csv"
res_file = "results.csv"

TROUSER = "trouser"
SHIRT = "tshirt-top"
LABEL = "label"

TROUSER_CAT = 0
SHIRT_CAT = 1
DI_TO_CAT = {TROUSER: TROUSER_CAT, SHIRT: SHIRT_CAT}
CAT_TO_DI = {SHIRT_CAT: SHIRT, TROUSER_CAT: TROUSER}

df = pd.read_csv(train_file)
df[LABEL] = pd.Categorical(df[LABEL].map(CAT_TO_DI))
display(df.head(10))

fig, ax = plt.subplots(3, 5, figsize=(10, 6))
for j in range(3):
    for i in range(5):
        ax[j, i].imshow(df.drop(LABEL, axis=1).iloc[5 * j + i].to_numpy().reshape(28, 28))
        ax[j, i].set_axis_off()
plt.show()

	label	...
0	tshirt-top	...
1	trouser	...
2	tshirt-top	...
3	trouser	...
4	tshirt-top	...
5	tshirt-top	...
6	trouser	...
7	tshirt-top	...
8	trouser	...
9	trouser	...

10 rows × 785 columns

png

Train-validation-test split

X_all = df.drop(LABEL, axis=1)
y_all = df.loc[:, LABEL]

X_train_val, X_test, y_train_val, y_test = train_test_split(
    X_all, y_all, test_size=0.2, random_state=RANDOM_STATE
)
X_train, X_val, y_train, y_val = train_test_split(
    X_train_val, y_train_val, test_size=0.25, random_state=RANDOM_STATE
)

d = pd.DataFrame(
    [
        ["train", X_train.shape[0], X_train.shape[0] / X_all.shape[0]],
        ["validation", X_val.shape[0], X_val.shape[0] / X_all.shape[0]],
        ["test", X_test.shape[0], X_test.shape[0] / X_all.shape[0]],
    ],
    columns=["dataset", "count", "relative count"],
)
display(d)

fig, ax = plt.subplots(figsize=(4, 4))
ax.pie(d["count"], labels=d["dataset"], autopct="%1.f%%")
ax.set_title("Train-test split")
plt.show()

	dataset	count	relative count
0	train	1440	0.6
1	validation	480	0.2
2	test	480	0.2

png

The dataset is split into training (1440 data points), validation (480 data points), and test sets (480 data points).

Exploratory Data Analysis

Features (Pixels)

d = pd.DataFrame(
    [
        ["count", X_train.shape[0]],
        ["size", X_train.shape[1]],
        ["min", X_train.stack().min()],
        ["mean", X_train.stack().mean().astype(int)],
        ["max", X_train.stack().max()],
    ],
    columns=["stat", "value"],
)
display(d)

d = X_train.to_numpy().flatten()
d2 = pd.DataFrame(
    [
        ["black pixel", d[d == 0].shape[0]],
        ["non-black pixel", d[d != 0].shape[0]],
    ],
    columns=["color", "count"],
)

fig, ax = plt.subplots(figsize=(6, 3))
ax.bar(d2["color"], d2["count"])
ax.set_title("Pixel count: black v. non-black")
ax.set_ylabel("count")
plt.show()

d = X_train.to_numpy().flatten()

fig, ax = plt.subplots(figsize=(6, 4))
ax.hist(d[d > 1], bins=20)
ax.set_title("Histogram of gray level distribution (non-black pixels)")
ax.set_xlabel("grey level")
ax.set_ylabel("frequency")
plt.show()

	stat	value
0	count	1440
1	size	784
2	min	0
3	mean	60
4	max	255

png

The data consists of 28x28 pixel images (784 features) in grayscale (values from 0 to 255). A significant portion of the image data is composed of black pixels (value 0).

Target Variable

X_shirts = X_train[y_train == SHIRT]
y_shirts = y_train[y_train == SHIRT]

X_trousers = X_train[y_train == TROUSER]
y_trousers = y_train[y_train == TROUSER]

d = pd.DataFrame(
    [["tshirt-top", y_shirts.shape[0]], ["trouser", y_trousers.shape[0]]],
    columns=["label", "count"],
)
display(d)

d = y_train.value_counts().to_frame()
fig, ax = plt.subplots(figsize=(4, 4))
ax.pie(d["count"], labels=d.index, autopct="%1.2f%%")
plt.show()

fig, ax = plt.subplots(2, 5, figsize=(10, 4))
fig.suptitle("T-shirts / Tops")
for j in range(2):
    for i in range(5):
        ax[j, i].imshow(X_shirts.iloc[5 * j + i].to_numpy().reshape(28, 28))
        ax[j, i].set_axis_off()
plt.show()

fig, ax = plt.subplots(2, 5, figsize=(10, 4))
fig.suptitle("Trousers")
for j in range(2):
    for i in range(5):
        ax[j, i].imshow(X_trousers.iloc[5 * j + i].to_numpy().reshape(28, 28))
        ax[j, i].set_axis_off()
plt.show()

	label	count
0	tshirt-top	687
1	trouser	753

png

The task is binary classification between T-shirts/tops and trousers. The classes are relatively balanced.

Model Development

Support Vector Machine (SVM)

SVM is a suitable model for this data:

It performs well in high-dimensional spaces.
It is relatively resistant to overfitting.
It can create non-linear decision boundaries.
Although computationally intensive, the dataset size is small enough that it won’t be an issue.

param_grid = ParameterGrid(
    {
        "kernel": ["linear", "poly", "rbf"],
        "C": list(chain(range(1, 15), range(15, 61, 5))),
        "preprocessing": [
            ("none", FunctionTransformer(lambda x: x)),
            ("normalized", Normalizer().fit(X_train)),
            ("standardized", StandardScaler().fit(X_train)),
        ],
    }
)

path = models / "svm.pickle"
if path.is_file():
    d = pickle.load(open(path, "rb"))
else:
    d = pd.DataFrame(columns=["preprocessing", "kernel", "C", "val_accuracy"])
    for i, params in enumerate(param_grid):
        k = params["kernel"]
        c = params["C"]
        p_name, p = params["preprocessing"]

        clf = SVC(kernel=k, C=c, cache_size=1000, random_state=RANDOM_STATE)
        clf.fit(p.transform(X_train), y_train)

        d.loc[i] = [p_name, k, c, clf.score(p.transform(X_val), y_val)]
    pickle.dump(d, open(path, "wb"))

d.sort_values("val_accuracy", ascending=False, ignore_index=True, inplace=True)
display(d.head(5))

d = d.sort_values("C")
fig, ax = plt.subplots(figsize=(10, 6))
for p, ls in zip(["none", "normalized", "standardized"], ["solid", "dotted", "dashed"]):
    for k in ["linear", "poly", "rbf"]:
        d1 = d.loc[(d["preprocessing"] == p) & (d["kernel"] == k)]
        ax.plot(d1["C"], d1["val_accuracy"], label=f"{p}, {k}", linestyle=ls)
ax.legend(loc="lower right", title="preprocessing, kernel", ncol=3)
ax.set_xlabel("parameter C")
ax.set_ylabel("validation accuracy")
ax.set_title("SVM learning curve")
plt.show()

	preprocessing	kernel	C	val_accuracy
0	none	rbf	25	0.9875
1	none	rbf	9	0.9875
2	none	rbf	55	0.9875
3	none	rbf	12	0.9875
4	none	rbf	35	0.9875

png

The SVM model achieves an accuracy of 98.75%. Normalization and standardization are particularly beneficial for linear and polynomial kernels. However, for the Gaussian (RBF) kernel, preprocessing does not significantly improve performance; in fact, the best result is achieved without it. The linear kernel is the weakest among the tested kernel functions.

Naive Bayes

Naive Bayes is suitable here:

It handles high-dimensional problems effectively by dealing well with irrelevant features.
The model is computationally efficient for this dataset.

param_grid = ParameterGrid(
    {
        "var_smoothing": np.linspace(1e-20, 3, num=100),
        "preprocessing": [
            ("none", FunctionTransformer(lambda x: x)),
            ("normalized", Normalizer().fit(X_train)),
            ("standardized", StandardScaler().fit(X_train)),
        ],
    }
)

path = models / "nb.pickle"
if path.is_file():
    d = pickle.load(open(path, "rb"))
else:
    d = pd.DataFrame(columns=["preprocessing", "var_smoothing", "val_accuracy"])
    for i, params in enumerate(param_grid):
        v = params["var_smoothing"]
        p_name, p = params["preprocessing"]

        clf = GaussianNB(var_smoothing=v)
        clf.fit(p.transform(X_train), y_train)

        d.loc[i] = [p_name, v, clf.score(p.transform(X_val), y_val)]
    pickle.dump(d, open(path, "wb"))

d.sort_values("val_accuracy", ascending=False, ignore_index=True, inplace=True)
display(d.head(5))

d = d.sort_values("var_smoothing")
fig, ax = plt.subplots(figsize=(10, 6))
for p, ls in zip(["none", "normalized", "standardized"], ["solid", "dashed", "dotted"]):
    d1 = d.loc[d["preprocessing"] == p]
    ax.plot(d1["var_smoothing"], d1["val_accuracy"], label=p, linestyle=ls)
ax.legend(loc="center right", title="preprocessing")
ax.set_xlabel("parameter var_smoothing")
ax.set_ylabel("validation accuracy")
ax.set_title("Naive Bayes learning curve")
plt.show()

	preprocessing	var_smoothing	val_accuracy
0	normalized	0.1818	0.9521
1	normalized	0.1515	0.9521
2	normalized	0.2424	0.9500
3	normalized	0.3030	0.9500
4	normalized	0.2727	0.9500

png

The Naive Bayes model achieves an accuracy of 95.21%. Tuning the var_smoothing parameter helps prevent issues with zero probabilities. Normalizing features improved the model’s performance in this case. However, it is weaker than the SVM model.

Linear Discriminant Analysis (LDA)

LDA is a reasonable choice here:

It is computationally efficient and interpretable.
It resists overfitting and performs well with smaller datasets.
Its linear decision boundary can be a limitation.
The model is sensitive to outliers.

param_grid = ParameterGrid(
    {
        "preprocessing": [
            ("none", FunctionTransformer(lambda x: x)),
            ("normalized", Normalizer().fit(X_train)),
            ("standardized", StandardScaler().fit(X_train)),
        ]
    }
)

path = models / "lda.pickle"
if path.is_file():
    d = pickle.load(open(path, "rb"))
else:
    d = pd.DataFrame(columns=["preprocessing", "val_accuracy"])
    for i, params in enumerate(param_grid):
        p_name, p = params["preprocessing"]

        clf = LinearDiscriminantAnalysis()
        clf.fit(p.transform(X_train), y_train)

        d.loc[i] = [p_name, clf.score(p.transform(X_val), y_val)]
    pickle.dump(d, open(path, "wb"))

d.sort_values("val_accuracy", ascending=False, ignore_index=True, inplace=True)
display(d)

fig, ax = plt.subplots(figsize=(6, 3))
for p in ["none", "normalized", "standardized"]:
    d1 = d.loc[d["preprocessing"] == p]
    ax.bar(p, d1["val_accuracy"], alpha=0.75)
ax.set_xlabel("preprocessing")
ax.set_ylabel("validation accuracy")
ax.set_title("Linear Discriminant Analysis")
plt.show()

	preprocessing	val_accuracy
0	none	0.9125
1	normalized	0.9125
2	standardized	0.9125

png

lda = LinearDiscriminantAnalysis()
X_lda = lda.fit_transform(Normalizer().fit_transform(X_train), y_train)

fig, ax = plt.subplots(figsize=(8, 2))
for y in [TROUSER, SHIRT]:
    x = X_lda[y_train == y]
    plt.scatter(x, np.zeros(x.shape[0]), label=y, alpha=0.4, marker="|", s=90)
ax.legend(loc="upper center", ncols=2)
ax.set_yticks([])
ax.set_xlabel("LDA component 1")
ax.set_title("Dimensionality reduction using LDA")
plt.show()

png

The LDA model achieves an accuracy of 91.25%. Neither normalization nor standardization affects the model’s performance. Training is straightforward with no hyperparameter tuning. Among the tested models, LDA is the weakest in terms of accuracy.

Modeling with PCA Dimensionality Reduction

pca_max = PCA()
scaler = Normalizer()
scaler.fit(X_train)
pca_max.fit(scaler.transform(X_train))

pcas = [PCA(n_components=i, random_state=RANDOM_STATE) for i in range(1, X_train.shape[1] + 1)]
for i, pca in enumerate(pcas):
    pca.components_ = pca_max.components_[: i + 1]
    pca.singular_values_ = pca_max.singular_values_[: i + 1]
    pca.mean_ = pca_max.mean_
    pca.n_components_ = i + 1

Projekce do 2D/3D prostoru

sample = np.random.choice(X_train.shape[0], 600, replace=False)
d = pcas[2].transform(scaler.transform(X_train.iloc[sample]))
d_trousers = d[y_train.iloc[sample] == TROUSER]
d_shirts = d[y_train.iloc[sample] == SHIRT]

fig, ax = plt.subplots(figsize=(14, 8))
ax.scatter(d_trousers[:, 0], d_trousers[:, 1], label="trouser")
ax.scatter(d_shirts[:, 0], d_shirts[:, 1], label="tshirt / top")
ax.legend(loc="upper right")
ax.set_xlabel("1. principal component")
ax.set_ylabel("2. principal component")
ax.set_xticklabels([])
ax.set_yticklabels([])
ax.set_title("2D projection of a random sample")
plt.show()

fig = plt.figure(figsize=(14, 18))
axes = []
for i in range(3):
    for j in range(2):
        ax = fig.add_subplot(3, 2, i * 2 + j + 1, projection="3d")
        ax.scatter(
            d_trousers[:, 0],
            d_trousers[:, 1],
            d_trousers[:, 2],
            label="trouser",
        )
        ax.scatter(d_shirts[:, 0], d_shirts[:, 1], d_shirts[:, 2], label="tshirt / top")
        # ax.legend(loc='upper right')
        ax.set_xlabel("1. component")
        ax.set_ylabel("2. component")
        ax.set_zlabel("3. component")
        ax.set_xticklabels([])
        ax.set_yticklabels([])
        ax.set_zticklabels([])
        axes.append(ax)
axes[0].view_init(20, 0)
axes[1].view_init(20, 30)
axes[2].view_init(20, 80)
axes[3].view_init(40, 80)
axes[4].view_init(60, 80)
axes[5].view_init(80, 80)
axes[0].legend(loc="upper left")
fig.suptitle("3D projection of a random sample", y=0.9)
plt.show()

png

The data points are relatively well-separated even with only 2 or 3 principal components.

SVM with PCA

param_grid = ParameterGrid(
    {
        "kernel": ["linear", "poly", "rbf"],
        "C": [9],
        "pca": list(
            chain(
                range(10),
                range(10, 100, 10),
                range(100, 200, 30),
                range(200, 501, 100),
            )
        ),
    }
)

path = models / "svm-pca.pickle"
if path.is_file():
    d = pickle.load(open(path, "rb"))
else:
    d = pd.DataFrame(columns=["kernel", "C", "PCA components", "val_accuracy"])
    for i, params in enumerate(param_grid):
        k = params["kernel"]
        c = params["C"]
        n = params["pca"]
        p = pcas[n]

        if k == "linear":
            clf = LinearSVC(C=c, dual="auto", random_state=RANDOM_STATE)
        else:
            clf = SVC(kernel=k, C=c, cache_size=1000, random_state=RANDOM_STATE)
        xin = p.transform(X_train.values)
        clf.fit(xin, y_train)

        d.loc[i] = [k, c, n, clf.score(p.transform(X_val.values), y_val)]
    pickle.dump(d, open(path, "wb"))

d.sort_values("val_accuracy", ascending=False, ignore_index=True, inplace=True)
display(d.head(5))

d = d.sort_values("PCA components")
fig, ax = plt.subplots(figsize=(10, 5))
for k, ls in zip(["linear", "poly", "rbf"], ["solid", "dotted", "dashed"]):
    d1 = d.loc[d["kernel"] == k]
    ax.plot(d1["PCA components"], d1["val_accuracy"], label=k, linestyle=ls)
ax.legend(loc="lower right", title="kernel")
ax.set_xlabel("number of principal components")
ax.set_ylabel("validation accuracy")
ax.set_title("SVM learning curve (with PCA dimensionality reduction)")
plt.show()

	kernel	C	PCA components	val_accuracy
0	rbf	9	140	0.9875
1	rbf	9	40	0.9854
2	rbf	9	500	0.9833
3	rbf	9	50	0.9833
4	rbf	9	400	0.9833

png

SVM with PCA reduction performs similarly to the original SVM without reduction. The data resides in a lower-dimensional subspace; the best model used 140 principal components, and the second best, with only a marginal difference in performance, used 40. This represents a significant dimensionality reduction from the original 784 features.

Naive Bayes with PCA

ll = np.linspace(1e-2, 10, num=10)
param_grid = ParameterGrid({"var_smoothing": ll, "pca": list(chain(range(10), range(10, 100, 10)))})


path = models / "nb-pca.pickle"
if path.is_file():
    d = pickle.load(open(path, "rb"))
else:
    d = pd.DataFrame(columns=["var_smoothing", "PCA components", "val_accuracy"])
    for i, params in enumerate(param_grid):
        v = params["var_smoothing"]
        n = params["pca"]
        p = pcas[n]

        clf = GaussianNB(var_smoothing=v)
        clf.fit(p.transform(X_train.values), y_train)

        d.loc[i] = [v, n, clf.score(p.transform(X_val.values), y_val)]
    pickle.dump(d, open(path, "wb"))

d.sort_values("val_accuracy", ascending=False, ignore_index=True, inplace=True)
display(d.head(5))

d = d.sort_values("PCA components")
fig, ax = plt.subplots(figsize=(10, 6))
for v in ll:
    d1 = d.loc[d["var_smoothing"] == v]
    ax.plot(d1["PCA components"], d1["val_accuracy"], label=f"{v:.2f}")
ax.legend(loc="upper right", title="var smoothing", ncols=2)
ax.set_xlabel("number of principal components")
ax.set_ylabel("validation accuracy")
ax.set_title("Naive Bayes learning curve (with PCA dimensionality reduction)")
plt.show()

	var_smoothing	PCA components	val_accuracy
0	1.12	6.0	0.9458
1	1.12	10.0	0.9458
2	1.12	9.0	0.9458
3	0.01	1.0	0.9437
4	2.23	4.0	0.9437

png

PCA dimensionality reduction improved Naive Bayes performance. The model performs best with 6 components. However, its peak performance is still lower than that of SVM.

LDA with PCA

param_grid = ParameterGrid(
    {
        "preprocessing": [
            ("none", FunctionTransformer(lambda x: x)),
            ("normalized", Normalizer().fit(X_train)),
        ],
        "pca": list(chain(range(1, 10), range(10, 101, 10))),
    }
)

path = models / "lda-pca.pickle"
if path.is_file():
    d = pickle.load(open(path, "rb"))
else:
    d = pd.DataFrame(columns=["preprocessing", "PCA components", "val_accuracy"])
    for i, params in enumerate(param_grid):
        n = params["pca"]
        pn, p = params["preprocessing"]

        clf = LinearDiscriminantAnalysis()
        x = p.transform(X_train)
        x = PCA(n, random_state=RANDOM_STATE).fit_transform(x)
        clf.fit(x, y_train)

        x = p.transform(X_val)
        x = PCA(n, random_state=RANDOM_STATE).fit_transform(x)
        d.loc[i] = [pn, n, clf.score(x, y_val)]
    pickle.dump(d, open(path, "wb"))

d.sort_values("val_accuracy", ascending=False, ignore_index=True, inplace=True)
display(d.head(5))

d = d.sort_values("PCA components")
fig, ax = plt.subplots(figsize=(10, 4))
for p, ls in zip(["none", "normalized"], ["solid", "dashed"]):
    d1 = d.loc[d["preprocessing"] == p]
    ax.plot(d1["PCA components"], d1["val_accuracy"], label=p, linestyle=ls)
ax.legend(loc="lower right", title="preprocessing")
ax.set_xlabel("number of principal components")
ax.set_ylabel("validation accuracy")
ax.set_title("LDA learning curve (with PCA dimensionality reduction)")
plt.show()

	preprocessing	PCA components	val_accuracy
0	normalized	10	0.9500
1	normalized	5	0.9500
2	normalized	9	0.9500
3	normalized	30	0.9500
4	normalized	20	0.9479

png

PCA reduction significantly improved LDA performance, increasing accuracy from 91% to 95% with only 5 principal components. Performance slightly decreases with more components.

Modeling with LLE Dimensionality Reduction

Non-linear Projection to 2D/3D Space

lle = LocallyLinearEmbedding(n_neighbors=5, n_components=3)
lle.fit(X_train)

np.random.seed(RANDOM_STATE)
sample = np.random.choice(X_train.shape[0], 600, replace=False)
xs, ys = X_train.iloc[sample], y_train.iloc[sample]
d = lle.transform(xs)
d_trousers = d[ys == TROUSER]
d_shirts = d[ys == SHIRT]

fig, ax = plt.subplots(figsize=(10, 8))
ax.scatter(d_trousers[:, 0], d_trousers[:, 1], label="trouser")
ax.scatter(d_shirts[:, 0], d_shirts[:, 1], label="tshirt / top")
ax.legend(loc="upper right")
ax.set_xlabel("1. LLE component")
ax.set_ylabel("2. LLE component")
ax.set_xticklabels([])
ax.set_yticklabels([])
ax.set_title("2D projection of a random sample")
plt.show()

fig = plt.figure(figsize=(14, 12))
axes = []
for i in range(2):
    for j in range(2):
        ax = fig.add_subplot(2, 2, i * 2 + j + 1, projection="3d")
        ax.scatter(
            d_trousers[:, 0],
            d_trousers[:, 1],
            d_trousers[:, 2],
            label="trouser",
        )
        ax.scatter(d_shirts[:, 0], d_shirts[:, 1], d_shirts[:, 2], label="tshirt / top")
        # ax.legend(loc='upper right')
        ax.set_xlabel("1. component")
        ax.set_ylabel("2. component")
        ax.set_zlabel("3. component")
        ax.set_xticklabels([])
        ax.set_yticklabels([])
        ax.set_zticklabels([])
        axes.append(ax)
axes[0].view_init(20, 60)
axes[1].view_init(60, 20)
axes[2].view_init(70, -40)
axes[3].view_init(90, -90)
axes[0].legend(loc="upper left")
fig.suptitle("3D projection of a random sample", y=0.9)
plt.show()

png

Similar to PCA, the data points are well-separated even with only 2 LLE components.

SVM with LLE

param_grid = ParameterGrid(
    {
        "kernel": ["linear", "poly", "rbf"],
        "C": [10],
        "lle": list(chain(range(1, 10), range(10, 30, 5), range(30, 100, 20))),
    }
)

path = models / "svm-lle.pickle"
if path.is_file():
    d = pickle.load(open(path, "rb"))
else:
    d = pd.DataFrame(columns=["kernel", "C", "LLE components", "val_accuracy"])
    for i, params in enumerate(param_grid):
        k = params["kernel"]
        c = params["C"]
        n = params["lle"]
        p = LocallyLinearEmbedding(n_neighbors=10, n_components=n)
        p.fit(X_train.values)

        if k == "linear":
            clf = LinearSVC(C=c, dual="auto", random_state=RANDOM_STATE)
        else:
            clf = SVC(kernel=k, C=c, cache_size=1000, random_state=RANDOM_STATE)
        xin = p.transform(X_train.values)
        clf.fit(xin, y_train)

        d.loc[i] = [k, c, n, clf.score(p.transform(X_val.values), y_val)]
    pickle.dump(d, open(path, "wb"))

d.sort_values("val_accuracy", ascending=False, ignore_index=True, inplace=True)
display(d.head(5))

d = d.sort_values("LLE components")
fig, ax = plt.subplots(figsize=(10, 5))
for k, ls in zip(["linear", "poly", "rbf"], ["solid", "dotted", "dashed"]):
    d1 = d.loc[d["kernel"] == k]
    ax.plot(d1["LLE components"], d1["val_accuracy"], label=k, linestyle=ls)
ax.legend(loc="lower left", title="kernel")
ax.set_xlabel("number of components")
ax.set_ylabel("validation accuracy")
ax.set_title("SVM learning curve (with LLE dimensionality reduction)")
plt.show()

	kernel	C	LLE components	val_accuracy
0	rbf	10	70	0.9729
1	rbf	10	25	0.9708
2	rbf	10	90	0.9688
3	rbf	10	50	0.9688
4	rbf	10	30	0.9688

png

LLE dimensionality reduction does not improve SVM performance. In fact, for a high number of components, such as with a polynomial kernel, the accuracy decreases.

Naive Bayes with LLE

ll = np.linspace(1e-1, 2, num=5)
param_grid = ParameterGrid(
    {"var_smoothing": ll, "lle": list(chain(range(1, 10), range(10, 100, 10)))}
)

path = models / "nb-lle.pickle"
if path.is_file():
    d = pickle.load(open(path, "rb"))
else:
    d = pd.DataFrame(columns=["var_smoothing", "LLE components", "val_accuracy"])
    for i, params in enumerate(param_grid):
        v = params["var_smoothing"]
        n = params["lle"]
        p = LocallyLinearEmbedding(n_neighbors=10, n_components=n)
        p.fit(X_train.values)

        clf = GaussianNB(var_smoothing=v)
        clf.fit(p.transform(X_train.values), y_train)

        d.loc[i] = [v, n, clf.score(p.transform(X_val.values), y_val)]
    pickle.dump(d, open(path, "wb"))

d.sort_values("val_accuracy", ascending=False, ignore_index=True, inplace=True)
display(d.head(5))

d = d.sort_values("LLE components")
fig, ax = plt.subplots(figsize=(10, 6))
for v in ll:
    d1 = d.loc[d["var_smoothing"] == v]
    ax.plot(d1["LLE components"], d1["val_accuracy"], label=f"{v:.2f}")
ax.legend(loc="lower left", title="var smoothing")
ax.set_xlabel("number of principal components")
ax.set_ylabel("validation accuracy")
ax.set_title("Naive Bayes learning curve (with LLE dimensionality reduction)")
plt.show()

	var_smoothing	LLE components	val_accuracy
0	0.100	10.0	0.9604
1	0.100	9.0	0.9604
2	0.100	8.0	0.9604
3	0.575	10.0	0.9583
4	0.575	9.0	0.9583

png

LLE mildly improves Naive Bayes. Dimensionality reduction results in the best Naive Bayes model to date.

LDA with LLE

param_grid = ParameterGrid(
    {
        "preprocessing": [
            ("none", FunctionTransformer(lambda x: x)),
            ("normalized", Normalizer().fit(X_train.values)),
        ],
        "lle": list(chain(range(1, 100, 20), range(100, 700, 100))),
    }
)

path = models / "lda-lle.pickle"
if path.is_file():
    d = pickle.load(open(path, "rb"))
else:
    d = pd.DataFrame(columns=["preprocessing", "LLE components", "val_accuracy"])
    for i, params in enumerate(param_grid):
        n = params["lle"]
        pn, p = params["preprocessing"]
        p2 = LocallyLinearEmbedding(n_neighbors=25, n_components=n)
        p2.fit(p.transform(X_train.values))

        clf = LinearDiscriminantAnalysis()
        x = p.transform(X_train.values)
        x = p2.transform(x)
        clf.fit(x, y_train)

        x = p.transform(X_val.values)
        x = p2.transform(x)
        d.loc[i] = [pn, n, clf.score(x, y_val)]
    pickle.dump(d, open(path, "wb"))

d.sort_values("val_accuracy", ascending=False, ignore_index=True, inplace=True)
display(d.head(5))

d = d.sort_values("LLE components")
fig, ax = plt.subplots(figsize=(10, 4))
for p, ls in zip(["none", "normalized"], ["solid", "dashed"]):
    d1 = d.loc[d["preprocessing"] == p]
    ax.plot(d1["LLE components"], d1["val_accuracy"], label=p, linestyle=ls)
ax.legend(loc="lower right", title="preprocessing")
ax.set_xlabel("number of principal components")
ax.set_ylabel("validation accuracy")
ax.set_title("LDA learning curve (with LLE dimensionality reduction)")
plt.show()

	preprocessing	LLE components	val_accuracy
0	normalized	400	0.9854
1	normalized	600	0.9833
2	normalized	500	0.9833
3	none	600	0.9812
4	none	500	0.9812

png

d = pd.DataFrame(
    [["none", 0.9125], ["PCA", 0.95], ["LLE", 0.9878]],
    columns=["dim. reduction", "val_accuracy"],
)
ticks = [d.iloc[0, 1], d.iloc[1, 1], d.iloc[2, 1]]

display(d)

fig, (ax1, ax2) = plt.subplots(2, 1, figsize=(6, 4))
fig.subplots_adjust(hspace=0.05)
for p in ["none", "PCA", "LLE"]:
    d1 = d.loc[d["dim. reduction"] == p]
    ax1.bar(p, d1["val_accuracy"], alpha=0.75)
    ax2.bar(p, d1["val_accuracy"], alpha=0.75)
ax1.set_ylim(0.85, 1)
ax2.set_ylim(0, 0.15)
ax1.spines.bottom.set_visible(False)
ax2.spines.top.set_visible(False)
ax1.set_xticks([])
ax1.set_yticks(ticks)
ax2.set_yticks([0, 0.04, 0.08, 0.12])
ax1.xaxis.tick_top()

for y in ticks:
    ax1.axhline(y, alpha=0.4, c="black", linestyle="--", linewidth=0.51)

d = 0.5
kwargs = dict(
    marker=[(-1, -d), (1, d)],
    markersize=12,
    linestyle="none",
    color="k",
    mec="k",
    mew=1,
    clip_on=False,
)
ax1.plot([0, 1], [0, 0], transform=ax1.transAxes, **kwargs)
ax2.plot([0, 1], [1, 1], transform=ax2.transAxes, **kwargs)

ax2.set_xlabel("dimentionality reduction")
ax2.set_ylabel("validation accuracy")
ax1.set_title("Linear Discriminant Analysis (dim. reduction)")
plt.show()

	dim. reduction	val_accuracy
0	none	0.9125
1	PCA	0.9500
2	LLE	0.9878

png

LLE dimensionality reduction significantly improved LDA performance. LDA with LLE now achieves performance comparable to the best SVM model found earlier. This improvement is likely due to LLE’s non-linear capabilities, allowing LDA to overcome its linear decision boundary limitations.

Final Model

Best Performing Model

The best model in terms of accuracy is LDA with LLE dimensionality reduction after normalization. This model will be retrained on the full training dataset and used for evaluation on the test data.

Final Evaluation

class Model(BaseEstimator, ClassifierMixin):
    def __init__(self):
        self.normalizer = Normalizer()
        self.lle = LocallyLinearEmbedding(n_neighbors=25, n_components=400)
        self.clf = LinearDiscriminantAnalysis(store_covariance=True)

    def fit(self, X, y):
        X, y = check_X_y(X, y)
        self.classes_ = unique_labels(y)
        self.X_ = X
        self.y_ = y

        X1 = self.normalizer.fit_transform(X)
        X2 = self.lle.fit_transform(X1)
        return self.clf.fit(X2, y)

    def transform_x_(self, X):
        X1 = self.normalizer.transform(X)
        X2 = self.lle.transform(X1)
        return X2

    def predict(self, X):
        check_is_fitted(self)
        X = check_array(X)
        X = self.transform_x_(X)
        return self.clf.predict(X)

    def predict_proba(self, X):
        check_is_fitted(self)
        X = check_array(X)
        X = self.transform_x_(X)
        return self.clf.predict_proba(X)


pathm = models / "final-m.pickle"
pathd = models / "final-d.pickle"
if pathm.is_file() and pathd.is_file():
    m = pickle.load(open(pathm, "rb"))
    d = pickle.load(open(pathd, "rb"))
else:
    m = Model()
    m.fit(X_train_val, y_train_val)

    d = classification_report(y_test, m.predict(X_test), output_dict=True)
    d = pd.DataFrame(d).drop(["weighted avg", "macro avg"], axis=1)

    pickle.dump(m, open(pathm, "wb"))
    pickle.dump(d, open(pathd, "wb"))

pd.set_option("display.precision", 2)
display(d)

fig, axes = plt.subplots(2, 2, figsize=(10, 9))
fig.suptitle("Evaluation of LDA with normalization and LLE (test data)", size=16)
fig.subplots_adjust(hspace=0.3)

ax = axes[0][0]
ConfusionMatrixDisplay.from_estimator(m, X_test, y_test, ax=ax, cmap="Blues", colorbar=False)
ax.set_title("Confusion matrix of LDA")

ax = axes[0][1]
ConfusionMatrixDisplay.from_estimator(
    m, X_test, y_test, ax=ax, cmap="Blues", normalize="all", colorbar=False
)
ax.set_title("Confusion matrix of LDA (normalized)")
ax.set_ylabel("")
ax.set_yticklabels([])

ax = axes[1][0]
RocCurveDisplay.from_estimator(m, X_test, y_test, name="LDA", ax=ax, plot_chance_level=True)
ax.set_title("ROC curve")

ax = axes[1][1]
PrecisionRecallDisplay.from_estimator(m, X_test, y_test, name="LDA", ax=ax)
ax.set_title("Precision-Recall curve")

plt.show()

	trouser	tshirt-top	accuracy
precision	1.00	0.96	0.98
recall	0.96	1.00	0.98
f1-score	0.98	0.98	0.98
support	245.00	235.00	0.98

png

Model Performance Estimate

We expect an accuracy of 98% on new, unseen data.

Evaluation

edf = pd.read_csv("evaluate.csv")
id = edf.loc[:, "ID"]
X_eval = edf.drop("ID", axis=1)

m.fit(X_all, y_all)
y_eval_hat = m.predict(X_eval)

d = pd.DataFrame()
d["ID"] = id
d["label"] = np.vectorize(DI_TO_CAT.get)(y_eval_hat)
d.to_csv(res_file, index=False)

fig, ax = plt.subplots(3, 5, figsize=(10, 6))
fig.suptitle("Prediction of evaluation.csv", size=18)
for j in range(3):
    for i in range(5):
        pos = 5 * j + i
        ax[j, i].imshow(edf.drop("ID", axis=1).iloc[pos].to_numpy().reshape(28, 28))
        ax[j, i].set_axis_off()
        ax[j, i].set_title(f"({pos}) {y_eval_hat[pos]}")
plt.show()

png

Approach#

Data#

Train-validation-test split#

Exploratory Data Analysis#

Features (Pixels)#

Target Variable#

Model Development#

Support Vector Machine (SVM)#

Naive Bayes#

Linear Discriminant Analysis (LDA)#

Modeling with PCA Dimensionality Reduction#

Projekce do 2D/3D prostoru#

SVM with PCA#

Naive Bayes with PCA#

LDA with PCA#

Modeling with LLE Dimensionality Reduction#

Non-linear Projection to 2D/3D Space#

SVM with LLE#

Naive Bayes with LLE#

LDA with LLE#

Final Model#

Best Performing Model#

Final Evaluation#

Model Performance Estimate#

Evaluation#