如何使用MongoDB开发一个简单的机器学习系统

sudo service mongodb start

mongod

mongo
use mydb

mongoimport --db mydb --collection flowers --type csv --headerline --file iris.csv

db.flowers.find()

db.flowers.find({ species: "setosa" })

db.flowers.find({ sepal_length: { $gt: 5.0, $lt: 6.0 } })

pip install pymongo

from pymongo import MongoClient

# 连接MongoDB数据库
client = MongoClient()
db = client.mydb

# 查询数据集
flowers = db.flowers.find()

# 打印结果
for flower in flowers:
    print(flower)

from pymongo import MongoClient

# 连接MongoDB数据库
client = MongoClient()
db = client.mydb

# 计算数据集的平均值
average_sepal_length = db.flowers.aggregate([
    { "$group": {
        "_id": None,
        "avg_sepal_length": { "$avg": "$sepal_length" }
    }}
])

# 打印平均值
for result in average_sepal_length:
    print(result["avg_sepal_length"])

from pymongo import MongoClient
from sklearn.linear_model import LoGISticRegression
import pickle

# 连接MongoDB数据库
client = MongoClient()
db = client.mydb

# 查询数据集
flowers = db.flowers.find()

# 准备数据集
X = []
y = []

for flower in flowers:
    X.append([flower["sepal_length"], flower["sepal_width"], flower["petal_length"], flower["petal_width"]])
    y.append(flower["species"])

# 训练模型
model = LogisticRegression()
model.fit(X, y)

# 保存模型
pickle.dump(model, open("model.pkl", "wb"))

# 加载模型
loaded_model = pickle.load(open("model.pkl", "rb"))

# 评估模型
accuracy = loaded_model.score(X, y)
print(accuracy)