当我们深入数据魔法的深处时，空气中弥漫着羊皮纸的气味和魔法能量的噼啪声。今天，我们将探索一种令人着迷的技术，称为One-Hot Encoding。想象一下，将一个顽皮的小精灵，以其充满活力的个性和不可预测的天性，转变为魔法地图上的一系列精确坐标。亲爱的读者，这就是One-Hot Encoding的本质。 ?♂️✨
正如麦格教授的魔杖可以将茶杯变成腊肠犬一样，这个咒语可以将分类数据（那些拒绝符合数值计算的讨厌的标签）转变为我们的模型可以理解的格式。可以将其视为将一群混乱的赫奇帕奇、拉文克劳、格兰芬多和斯莱特林变成由 1 和 0 组成的整齐网格，每个网格代表一个特定的学院。 ?✨
通过 One-Hot 编码，我们为每个独特的类别创建新列，并用 1 和 0 填充它们以指示存在或不存在。这就像将顽皮的家养小精灵分类到指定的房间中，确保每个小精灵都有自己的空间。到本章结束时，您将能够像经验丰富的魔咒大师一样充满信心地施展该咒语，将您的数据从纠结的结转变为组织精美的挂毯。 ?✨
在数据科学的迷人领域，数字起舞和模式显露出来，我们遇到了一种奇怪的信息，称为分类数据。与数字对应不同，这些数据点并不代表数量，而是代表不同的类别或组。
想象一下霍格沃茨的分院帽，这是一个古老的智慧魔法物品，可以将学生分配到他们正确的学院。格兰芬多学院、赫奇帕奇学院、拉文克劳学院和斯莱特林学院都是分类数据的示例。它们代表着具有独特特征的不同群体，就像房屋本身一样。同样，学生选择的宠物类型——忠诚的猫头鹰、咕噜咕噜的猫或脾气暴躁的蟾蜍——也属于分类数据的范畴。
分类数据就像在物体上放置神奇的标签，帮助我们区分和分类它们。正如草药学学生会根据不同植物的特性对其进行仔细分类一样，我们使用分类数据来排序和理解数据集中的不同元素。通过理解这些神奇的标签，我们可以解锁隐藏的模式并施展强大的咒语（分析）来揭开数据的秘密。让我们寻求一些进一步的知识，房子柱下面有什么价值，亲爱的巫师们，请施展你的魔杖。 ?✨
# Displaying the unique categories in the 'House' column
unique_houses = hogwarts_df['house'].unique()
print(f"Unique Houses: {unique_houses}")
Unique Houses: ['Gryffindor' 'Slytherin' 'Ravenclaw' 'Hufflepuff' 'Durmstrang' 'Beauxbatons']
理解这些类别至关重要，因为我们神奇的算法（或模型）需要知道如何解释这些数据。然而，这些算法常常难以处理非数字数据，因为它们更适合处理数字。这就是 One-Hot Encoding 的魔力发挥作用的地方。
想象一下我们的霍格沃茨学生记录，里面充满了迷人的细节，比如房子、魔杖类型和最喜欢的科目。这些品质就像神奇的符文一样，承载着独特的能量。然而，我们出色的数据模型虽然能够实现惊人的壮举，但无法直接破译这些印记。我们必须将它们转化为它们能够理解的语言——数字。
输入One-Hot Encoding咒语，这是一个强大的咒语，揭示了每个分类变量隐藏的本质。这就像在隐藏的密室上施展Lumos咒语，照亮每一个角落和缝隙。只需轻轻一挥编码棒，我们就能将每个类别转换为自己的独立列。例如，如果一名学生属于格兰芬多，那么 1 将神奇地出现在格兰芬多栏中，而其他学院栏则保持黑色。
这种转换类似于创建一个神奇的挂毯，其中每个线程代表一个类别。通过将这些线索编织在一起，我们创建了丰富而详细的学生画像，准备好通过我们的数据模型进行分析。就好像我们赋予我们的模型通过复方果汁饮用者的眼睛看世界的能力，体验每个学生的独特视角。让我们继续尝试对我们的第一列或功能进行 One-Hot 编码，首先尝试性别列。
# Importing necessary libraries
import pandas as pd
from sklearn.preprocessing import OneHotEncoder
from IPython.display import display, HTML
# Assuming hogwarts_df is already defined and contains the 'gender' column
# Applying One-Hot Encoding to the 'gender' column
encoder = OneHotEncoder(sparse_output=False) # Updated parameter name
encoded_data = encoder.fit_transform(hogwarts_df[['gender']])
# Converting the encoded data into a DataFrame and attaching it to the original dataset
encoded_df = pd.DataFrame(encoded_data, columns=encoder.get_feature_names_out(['gender']))
hogwarts_df = pd.concat([hogwarts_df, encoded_df], axis=1)
# Dropping the original 'gender' column as it's now encoded
hogwarts_df.drop('gender', axis=1, inplace=True)
# Displaying the transformed DataFrame in a scrollable pane
html = hogwarts_df.head(5).to_html() # Convert DataFrame to HTML
scrollable_html = f"""
<div style="height: 300px; overflow: auto;">
{html}
</div>
"""
display(HTML(scrollable_html))
name age origin specialty house blood_status pet wand_type patronus quidditch_position boggart favorite_class house_points gender_Female gender_Male 0 Harry Potter 11 England Defense Against the Dark Arts Gryffindor Half-blood Owl Holly Stag Seeker Dementor Defense Against the Dark Arts 150.0 0.0 1.0 1 Hermione Granger 11 England Transfiguration Gryffindor Muggle-born Cat Vine Otter Seeker Failure Arithmancy 200.0 1.0 0.0 2 Ron Weasley 11 England Chess Gryffindor Pure-blood Rat Ash Jack Russell Terrier Keeper Spider Charms 50.0 0.0 1.0 3 Draco Malfoy 11 England Potions Slytherin Pure-blood Owl Hawthorn Non-corporeal Seeker Lord Voldemort Potions 100.0 0.0 1.0 4 Luna Lovegood 11 Ireland Creatures Ravenclaw Half-blood Owl Fir Hare Seeker Her mother Creatures 120.0 1.0 0.0
And if you scroll to the right, you might notice that the dataset now has additional two columns, the gender_Female and the gender_Male on top of the existing one, while dropping the original gender column that was there previously.
In the grand tapestry of the wizarding world of data, there exist two primary categories of magical artifacts: Structured Data and Unstructured Data. These are the building blocks of our enchanting spells and powerful potions.
Categorical Data is akin to a well-organized Herbology garden, where every plant (data point) has its rightful place. It's like a neatly filled Hogwarts student record, with columns for names, houses, and wand types, all aligned in perfect order. Structured data is a wizard's delight, easily understood and manipulated with a flick of the wand (or a few lines of code). ?✨
On the other hand, Numerical Data is a sprawling Forbidden Forest, filled with magical creatures (data points) roaming freely. It's like a collection of owls' letters, each with its own unique style and format. This data can be as diverse as the stars in the night sky, ranging from social media posts to news articles, images, and even spoken words. While it holds immense potential, taming this wild magic requires special spells and a keen eye for patterns. ??
Categorical Data (Qualitative Data):
Numerical Data (Quantitative Data):
Imagine a world where numbers and words could converse, where the language of magic flowed seamlessly from one to the other. This is the realm of one-hot encoding, a powerful spell that transforms the mysterious world of text into the concrete world of numbers.
Just as a skilled Herbologist categorizes plants by their properties, one-hot encoding sorts textual data into distinct categories. Consider the Sorting Hat, which assigns students to houses based on their unique qualities. Similarly, one-hot encoding creates separate columns for each category, with values of 0 or 1 indicating whether a data point belongs to that category or not.
For instance, if you have a column representing the houses of Hogwarts students (Gryffindor, Slytherin, Ravenclaw, and Hufflepuff), one-hot encoding would conjure four new columns, one for each house. A value of 1 in the Gryffindor column would signify a student belonging to that house, while other columns would be filled with 0s. This numerical representation allows our magical models to understand and process textual information with ease. With that being said, let's try to transform the remaining of the columns in our dataset, and see what we have to work on. ?
# Importing necessary libraries
import pandas as pd
from sklearn.preprocessing import OneHotEncoder
from IPython.display import display, HTML
# Assuming hogwarts_df is already defined and contains the necessary columns
columns_to_encode = [
'origin', 'specialty', 'blood_status', 'pet', 'wand_type', 'patronus', 'quidditch_position', 'boggart', 'favorite_class'
]
# Creating an instance of OneHotEncoder
encoder = OneHotEncoder(sparse_output=False)
# List to hold encoded DataFrames
encoded_dfs = []
# Applying One-Hot Encoding to each column and storing the result
for column in columns_to_encode:
encoded_data = encoder.fit_transform(hogwarts_df[[column]])
encoded_df = pd.DataFrame(encoded_data, columns=encoder.get_feature_names_out([column]))
encoded_dfs.append(encoded_df)
# Concatenating all encoded DataFrames into one
encoded_df_combined = pd.concat(encoded_dfs, axis=1)
# Concatenating the encoded DataFrame with the original DataFrame
hogwarts_df = pd.concat([hogwarts_df, encoded_df_combined], axis=1)
# Dropping the original columns that were encoded
hogwarts_df.drop(columns=columns_to_encode, inplace=True)
# Displaying the transformed DataFrame in a scrollable pane
html = hogwarts_df.head(5).to_html() # Convert DataFrame to HTML
scrollable_html = f"""
<div style="height: 300px; overflow: auto;">
{html}
</div>
"""
display(HTML(scrollable_html))
name age house house_points gender_Female gender_Male origin_Bulgaria origin_England origin_Europe origin_France origin_Indonesia origin_Ireland origin_Scotland origin_USA origin_Wales specialty_Auror specialty_Baking specialty_Charms specialty_Chess specialty_Creatures specialty_Dark Arts specialty_Defense Against the Dark Arts specialty_Dueling specialty_Goat Charming specialty_Gossip specialty_Herbology specialty_History of Magic specialty_Household Charms specialty_Legilimency specialty_Magical Creatures specialty_Memory Charms specialty_Metamorphmagus specialty_Muggle Artifacts specialty_Obscurus specialty_Potions specialty_Quidditch specialty_Strength specialty_Transfiguration specialty_Transformation blood_status_Half-blood blood_status_Muggle-born blood_status_No-mag blood_status_Pure-blood pet_Cat pet_Demiguise pet_Dog pet_Goat pet_Owl pet_Phoenix pet_Rat pet_Snake pet_Toad wand_type_Alder wand_type_Ash wand_type_Birch wand_type_Blackthorn wand_type_Cedar wand_type_Cherry wand_type_Chestnut wand_type_Cypress wand_type_Ebony wand_type_Elder wand_type_Elm wand_type_Fir wand_type_Hawthorn wand_type_Hazel wand_type_Hemlock wand_type_Holly wand_type_Hornbeam wand_type_Maple wand_type_Oak wand_type_Pine wand_type_Rosewood wand_type_Rowan wand_type_Sword wand_type_Teak wand_type_Vine wand_type_Walnut wand_type_Willow wand_type_Yew patronus_Cat patronus_Doe patronus_Dog patronus_Eagle patronus_Hare patronus_Horse patronus_Jack Russell Terrier patronus_Lion patronus_Non-corporeal patronus_Otter patronus_Phoenix patronus_Serpent patronus_Stag patronus_Swan patronus_Wolf quidditch_position_Azkaban quidditch_position_Beater quidditch_position_Chaser quidditch_position_Keeper quidditch_position_Seeker boggart_Ariana's death boggart_Dementor boggart_Dueling boggart_Failure boggart_Full Moon boggart_Her mother boggart_Lily Potter boggart_Lord Voldemort boggart_Severus Snape boggart_Spider boggart_Tom Riddle favorite_class_Arithmancy favorite_class_Baking favorite_class_Charms favorite_class_Creatures favorite_class_Dark Arts favorite_class_Defense Against the Dark Arts favorite_class_Dueling favorite_class_Goat Charming favorite_class_Gossip favorite_class_Herbology favorite_class_Household Charms favorite_class_Legilimency favorite_class_Memory Charms favorite_class_Muggle Studies favorite_class_Obscurus favorite_class_Potions favorite_class_Quidditch favorite_class_Strength favorite_class_Transfiguration favorite_class_Transformation 0 Harry Potter 11 Gryffindor 150.0 0.0 1.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1 Hermione Granger 11 Gryffindor 200.0 1.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 1.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 2 Ron Weasley 11 Gryffindor 50.0 0.0 1.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 3 Draco Malfoy 11 Slytherin 100.0 0.0 1.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 4 Luna Lovegood 11 Ravenclaw 120.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
In the magical realm of data science, where numbers hold secrets and patterns dance in the shadows, there exists a particularly enchanting spell: one-hot encoding. This spell is a transfiguration charm, capable of transforming seemingly ordinary text into a numerical language that our magical computers can understand. Imagine a bustling Diagon Alley, filled with shops selling wands, cauldrons, and robes of every color. Each shop has a unique name, a string of letters that defines its identity. Now, picture these shop names as magical creatures, wild and untamed. To harness their power for our spells, we must transform them into something more manageable – numbers.
One-hot encoding is the spell that accomplishes this feat. It takes each unique shop name and creates a separate magical dimension (column) for it. Within these dimensions, we cast a binary spell, assigning a value of 1 to the shop that exists in that dimension and 0 to all others. It's like creating a magical grid, where each shop has its own spotlight moment. With this transformation, our once chaotic collection of shop names becomes an orderly array of numbers, ready to be analyzed and explored.?✨
To convert a numerical column with values ranging from 100 to 200 into a more machine learning-friendly format using one-hot encoding, you typically need to discretize the numerical values into categorical bins first. Here’s how to do it step by step:
You can create bins (categories) for the numerical values. For example, you might define bins like this:
Next, assign each numerical value to its corresponding bin. For example:
| Original Value | Category |
|---|---|
| 100 | 100-120 |
| 110 | 100-120 |
| 125 | 121-140 |
| 145 | 141-160 |
| 165 | 161-180 |
| 180 | 161-180 |
| 200 | 181-200 |
Now, you can apply one-hot encoding to the categorical column. Each category will be represented as a binary vector:
| Original Value | 100-120 | 121-140 | 141-160 | 161-180 | 181-200 |
|---|---|---|---|---|---|
| 100 | 1 | 0 | 0 | 0 | 0 |
| 110 | 1 | 0 | 0 | 0 | 0 |
| 125 | 0 | 1 | 0 | 0 | 0 |
| 145 | 0 | 0 | 1 | 0 | 0 |
| 165 | 0 | 0 | 0 | 1 | 0 |
| 180 | 0 | 0 | 0 | 1 | 0 |
| 200 | 0 | 0 | 0 | 0 | 1 |
By discretizing the numerical values into categories and applying one-hot encoding, you transform the original numerical data into a format that machine learning algorithms can process effectively. This method captures the categorical nature of the data while preserving the information contained in the original numerical values.
import pandas as pd
from sklearn.preprocessing import OneHotEncoder
from IPython.display import display, HTML
# Step 1: Define bins and labels
bins = [100, 120, 140, 160, 180, 200] # Define the bin edges
labels = ['hp_100_120', 'hp_121_140', 'hp_141_160', 'hp_161_180', 'hp_181_200'] # Define the bin labels
# Step 2: Create a new categorical column based on the bins
hogwarts_df['house_category'] = pd.cut(hogwarts_df['house_points'], bins=bins, labels=labels, right=True)
# Step 3: One-hot encode the categorical column
hogwarts_df_encoded = pd.get_dummies(hogwarts_df, columns=['house_category'], prefix='', prefix_sep='')
# Replace True with 1 and False with 0
hogwarts_df_encoded = hogwarts_df_encoded.replace({True: 1, False: 0})
# Drop the house_points column
hogwarts_df_encoded.drop('house_points', axis=1, inplace=True)
# Displaying the transformed DataFrame in a scrollable pane
html = hogwarts_df_encoded.head(5).to_html() # Convert DataFrame to HTML & # Display first 5 rows in a scrollable pane
scrollable_html = f"""
<div style="height: 300px; overflow: auto;">
{html}
</div>
"""
display(HTML(scrollable_html))
name age house gender_Female gender_Male origin_Bulgaria origin_England origin_Europe origin_France origin_Indonesia origin_Ireland origin_Scotland origin_USA origin_Wales specialty_Auror specialty_Baking specialty_Charms specialty_Chess specialty_Creatures specialty_Dark Arts specialty_Defense Against the Dark Arts specialty_Dueling specialty_Goat Charming specialty_Gossip specialty_Herbology specialty_History of Magic specialty_Household Charms specialty_Legilimency specialty_Magical Creatures specialty_Memory Charms specialty_Metamorphmagus specialty_Muggle Artifacts specialty_Obscurus specialty_Potions specialty_Quidditch specialty_Strength specialty_Transfiguration specialty_Transformation blood_status_Half-blood blood_status_Muggle-born blood_status_No-mag blood_status_Pure-blood pet_Cat pet_Demiguise pet_Dog pet_Goat pet_Owl pet_Phoenix pet_Rat pet_Snake pet_Toad wand_type_Alder wand_type_Ash wand_type_Birch wand_type_Blackthorn wand_type_Cedar wand_type_Cherry wand_type_Chestnut wand_type_Cypress wand_type_Ebony wand_type_Elder wand_type_Elm wand_type_Fir wand_type_Hawthorn wand_type_Hazel wand_type_Hemlock wand_type_Holly wand_type_Hornbeam wand_type_Maple wand_type_Oak wand_type_Pine wand_type_Rosewood wand_type_Rowan wand_type_Sword wand_type_Teak wand_type_Vine wand_type_Walnut wand_type_Willow wand_type_Yew patronus_Cat patronus_Doe patronus_Dog patronus_Eagle patronus_Hare patronus_Horse patronus_Jack Russell Terrier patronus_Lion patronus_Non-corporeal patronus_Otter patronus_Phoenix patronus_Serpent patronus_Stag patronus_Swan patronus_Wolf quidditch_position_Azkaban quidditch_position_Beater quidditch_position_Chaser quidditch_position_Keeper quidditch_position_Seeker boggart_Ariana's death boggart_Dementor boggart_Dueling boggart_Failure boggart_Full Moon boggart_Her mother boggart_Lily Potter boggart_Lord Voldemort boggart_Severus Snape boggart_Spider boggart_Tom Riddle favorite_class_Arithmancy favorite_class_Baking favorite_class_Charms favorite_class_Creatures favorite_class_Dark Arts favorite_class_Defense Against the Dark Arts favorite_class_Dueling favorite_class_Goat Charming favorite_class_Gossip favorite_class_Herbology favorite_class_Household Charms favorite_class_Legilimency favorite_class_Memory Charms favorite_class_Muggle Studies favorite_class_Obscurus favorite_class_Potions favorite_class_Quidditch favorite_class_Strength favorite_class_Transfiguration favorite_class_Transformation hp_100_120 hp_121_140 hp_141_160 hp_161_180 hp_181_200 0 Harry Potter 11 Gryffindor 0.0 1.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0 0 1 0 0 1 Hermione Granger 11 Gryffindor 1.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 1.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0 0 0 0 1 2 Ron Weasley 11 Gryffindor 0.0 1.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0 0 0 0 0 3 Draco Malfoy 11 Slytherin 0.0 1.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0 0 0 0 0 4 Luna Lovegood 11 Ravenclaw 1.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1 0 0 0 0
Imagine a bustling Hogwarts Express, filled with students of all ages. From wide-eyed first-years to wise-cracking seventh-years, each student possesses a unique blend of magic and mischief. To better understand these young wizards and witches, we can transform their ages from precise numbers into magical categories.
This process, known as discretization, is like sorting mischievous house-elves into their designated chores. Instead of treating age as a continuous spectrum, we group students into specific age ranges or bins. Picture these bins as enchanted compartments on the Hogwarts Express, each holding students of similar age.
By discretizing age, we unlock a new dimension of magical insight. We can explore how different age groups perform in classes, participate in Quidditch, or even their preferred wand type. It's like casting a revealing charm on our data, uncovering hidden patterns and trends that would otherwise remain obscured. So, let's transform those numbers into magical categories and discover the enchanting secrets hidden within the age of our Hogwarts students! ?✨
import pandas as pd
from IPython.display import display, HTML
# Step 1: Define bins and labels for the age column
bins = [10, 12, 14, 16, 18] # Define the bin edges
labels = ['age_11', 'age_12', 'age_13', 'age_14'] # Define the bin labels
# Step 2: Create a new categorical column based on the bins
hogwarts_df_encoded['age_category'] = pd.cut(hogwarts_df_encoded['age'], bins=bins, labels=labels, right=True)
# Step 3: One-hot encode the categorical column
hogwarts_df_encoded_age = pd.get_dummies(hogwarts_df_encoded, columns=['age_category'], prefix='', prefix_sep='')
# Replace True with 1 and False with 0 (not necessary here since get_dummies already returns integers)
hogwarts_df_encoded_age = hogwarts_df_encoded_age.replace({True: 1, False: 0})
# Drop the age column
hogwarts_df_encoded_age.drop('age', axis=1, inplace=True)
# Displaying the transformed DataFrame in a scrollable pane
html = hogwarts_df_encoded_age.head(5).to_html() # Convert DataFrame to HTML & Display first 5 rows in a scrollable pane
scrollable_html = f"""
<div style="height: 300px; overflow: auto;">
{html}
</div>
"""
display(HTML(scrollable_html))
Imagine a bustling Hogwarts library, books overflowing from shelves, each with its own peculiar title. Without a proper cataloguing system, finding a specific spellbook would be like searching for a needle in a haystack, wouldn't it? In the realm of data, our columns are like these books. They hold valuable information, but without a clear and consistent naming system, they can quickly become a chaotic mess. This is where the magic of naming conventions comes into play.
Just as a librarian organizes books by subject, author, and title, we must impose order upon our columns. A well-crafted naming convention is like a powerful sorting spell, grouping similar columns together and making them easily identifiable. For example, using prefixes like 'student_', 'subject_', or 'score_' can instantly clarify the column's purpose.
By adopting a clear and consistent naming convention, you'll transform your data from a chaotic jumble into a well-organized magical library. This not only improves readability but also streamlines data manipulation and analysis. So, next time you're wrangling with data, remember the importance of a strong naming convention. It's the first step towards unlocking the hidden treasures within your dataset! ✨
import pandas as pd
from IPython.display import display, HTML
# Assuming hogwarts_df_encoded is already defined and contains the necessary columns
# Manipulate column names
hogwarts_df_encoded_age.columns = hogwarts_df_encoded_age.columns.str.lower().str.replace(' ', '_').str.replace('-', '-').str.replace("'", "")
# Display the transformed DataFrame in a scrollable pane
html = hogwarts_df_encoded_age.head(5).to_html() # Convert first 5 rows to HTML
scrollable_html = f"""
<div style="height: 300px; overflow: auto;">
{html}
</div>
"""
display(HTML(scrollable_html)) # Display first 5 rows in a scrollable pane
name house gender_female gender_male origin_bulgaria origin_england origin_europe origin_france origin_indonesia origin_ireland origin_scotland origin_usa origin_wales specialty_auror specialty_baking specialty_charms specialty_chess specialty_creatures specialty_dark_arts specialty_defense_against_the_dark_arts specialty_dueling specialty_goat_charming specialty_gossip specialty_herbology specialty_history_of_magic specialty_household_charms specialty_legilimency specialty_magical_creatures specialty_memory_charms specialty_metamorphmagus specialty_muggle_artifacts specialty_obscurus specialty_potions specialty_quidditch specialty_strength specialty_transfiguration specialty_transformation blood_status_half-blood blood_status_muggle-born blood_status_no-mag blood_status_pure-blood pet_cat pet_demiguise pet_dog pet_goat pet_owl pet_phoenix pet_rat pet_snake pet_toad wand_type_alder wand_type_ash wand_type_birch wand_type_blackthorn wand_type_cedar wand_type_cherry wand_type_chestnut wand_type_cypress wand_type_ebony wand_type_elder wand_type_elm wand_type_fir wand_type_hawthorn wand_type_hazel wand_type_hemlock wand_type_holly wand_type_hornbeam wand_type_maple wand_type_oak wand_type_pine wand_type_rosewood wand_type_rowan wand_type_sword wand_type_teak wand_type_vine wand_type_walnut wand_type_willow wand_type_yew patronus_cat patronus_doe patronus_dog patronus_eagle patronus_hare patronus_horse patronus_jack_russell_terrier patronus_lion patronus_non-corporeal patronus_otter patronus_phoenix patronus_serpent patronus_stag patronus_swan patronus_wolf quidditch_position_azkaban quidditch_position_beater quidditch_position_chaser quidditch_position_keeper quidditch_position_seeker boggart_arianas_death boggart_dementor boggart_dueling boggart_failure boggart_full_moon boggart_her_mother boggart_lily_potter boggart_lord_voldemort boggart_severus_snape boggart_spider boggart_tom_riddle favorite_class_arithmancy favorite_class_baking favorite_class_charms favorite_class_creatures favorite_class_dark_arts favorite_class_defense_against_the_dark_arts favorite_class_dueling favorite_class_goat_charming favorite_class_gossip favorite_class_herbology favorite_class_household_charms favorite_class_legilimency favorite_class_memory_charms favorite_class_muggle_studies favorite_class_obscurus favorite_class_potions favorite_class_quidditch favorite_class_strength favorite_class_transfiguration favorite_class_transformation hp_100_120 hp_121_140 hp_141_160 hp_161_180 hp_181_200 age_11 age_12 age_13 age_14 0 Harry Potter Gryffindor 0.0 1.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0 0 1 0 0 1 0 0 0 1 Hermione Granger Gryffindor 1.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 1.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0 0 0 0 1 1 0 0 0 2 Ron Weasley Gryffindor 0.0 1.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0 0 0 0 0 1 0 0 0 3 Draco Malfoy Slytherin 0.0 1.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0 0 0 0 0 1 0 0 0 4 Luna Lovegood Ravenclaw 1.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1 0 0 0 0 1 0 0 0
And once we've satisfied with the results, let's go ahead and save the current dataset to make the next enchanting journey more easier to navigate.
hogwarts_df_encoded_age.to_csv('data/hogwarts-students-03.csv', index=False)
Imagine a vast enchanted forest, teeming with magical creatures and extraordinary plants. Each creature and plant possesses unique qualities, and some may share hidden connections. To unveil these mysterious bonds, we must cast a spell of correlation. A correlation matrix is like a magical map, guiding us through this enchanted forest. Each point on the map represents a different creature or plant, and the lines connecting them reveal the strength of their relationship. A thick, vibrant line signifies a strong connection, while a thin, faint line indicates a weaker bond.
As we explore this magical map, we search for patterns and trends. Are certain creatures often found near specific plants? Do particular plants thrive in the company of others? By deciphering these connections, we can uncover hidden knowledge about the forest and its inhabitants. Just as a skilled Herbologist studies the interactions between plants, we, as data wizards, uncover the hidden relationships between variables. With the correlation matrix as our guide, we embark on a thrilling adventure to explore the magical tapestry of our data! ?️✨
# Importing necessary libraries
import pandas as pd
from IPython.display import display, HTML
# Assuming hogwarts_df is already defined and contains the necessary columns
# Selecting only numerical columns
numerical_df = hogwarts_df_encoded_age.select_dtypes(include=['number'])
# Calculating the correlation matrix
correlation_matrix = numerical_df.corr()
# Displaying the correlation matrix in a scrollable pane
correlation_html = correlation_matrix.to_html() # Convert correlation matrix to HTML
scrollable_correlation_html = f"""
<div style="height: 300px; overflow: auto;">
{correlation_html}
</div>
"""
display(HTML(scrollable_correlation_html)) # Display correlation matrix in a scrollable pane
And to make our analysis easier, let's just go ahead and save the correlation matrix into a tabular format.
correlation_matrix.to_csv('data/correlation-matrix.csv')
Just as a skilled potioneer carefully examines their ingredients before brewing a powerful concoction, we data scientists must meticulously inspect our data types. These data types are like the magical properties of our ingredients, determining how they will react in our spells (algorithms).
Imagine our dataset as a cauldron brimming with magical elements. Each element, be it a student's age, house, or wand type, has a specific form or essence - its data type. These types can be as diverse as the magical creatures of the Forbidden Forest: numbers (like the count of house points), text (like a student's name), dates (like the founding year of Hogwarts), and more.
Mismatched data types can lead to disastrous results, like a potion exploding or a spell backfiring. That's why we must cast a discerning eye over our data, ensuring each element is of the correct type. It's like checking if a Mandrake root is truly a root and not a disguised Goblin! Only then can we confidently proceed with our magical data transformations. ?♀️✨
# Importing necessary libraries
import pandas as pd
from IPython.display import display, HTML
# Assuming hogwarts_df is already defined and contains the necessary columns
# Setting display options to show all columns and prevent truncation
pd.set_option('display.max_columns', None) # Show all columns
pd.set_option('display.expand_frame_repr', False) # Prevent truncation in output
# Checking the data types of each column
data_types_df = hogwarts_df_encoded_age.dtypes.to_frame(name='Data Type') # Convert data types to a DataFrame
# Displaying the data types in a scrollable pane
data_types_html = data_types_df.to_html() # Convert DataFrame to HTML
scrollable_data_types_html = f"""
<div style="height: 150px; overflow: auto;">
{data_types_html}
</div>
"""
display(HTML(scrollable_data_types_html)) # Display data types in a scrollable pane
Data Type name object house object gender_female float64 gender_male float64 origin_bulgaria float64 origin_england float64 origin_europe float64 origin_france float64 origin_indonesia float64 origin_ireland float64 origin_scotland float64 origin_usa float64 origin_wales float64 specialty_auror float64 specialty_baking float64 specialty_charms float64 specialty_chess float64 specialty_creatures float64 specialty_dark_arts float64 specialty_defense_against_the_dark_arts float64 specialty_dueling float64 specialty_goat_charming float64 specialty_gossip float64 specialty_herbology float64 specialty_history_of_magic float64 specialty_household_charms float64 specialty_legilimency float64 specialty_magical_creatures float64
And now, dear apprentices, Gemika Haziq Nugroho has prepared a quiz to test your understanding of this powerful encoding spell. Are you ready to decode the mysteries of One-Hot Encoding?
Answer these questions to prove your mastery over the art of data transformation. As we continue our journey through the magical world of data science, remember that each spell we learn brings us closer to unveiling the secrets of our enchanted dataset. ??♂️✨
With these newfound skills, you're well-equipped to handle categorical data in any dataset. The path to becoming a master data wizard is full of wonder and discovery, and with each step, we draw closer to the heart of the magical data that surrounds us. Let us press on, eager to learn and ready to explore! ??
以上是Gemika 使用决策树算法对霍格沃茨学生进行排序的神奇指南（第 7 部分）的详细内容。更多信息请关注PHP中文网其他相关文章！
