To set a machine study pipe in the Google Cloud platform

Photo for Editor | Chatgt

Obvious Introduction

Machine study has been an important part of many companies, and businesses that do not use the risk of the back. As given to how serious models provide competitive benefits, it is only natural that many companies want to associate with their programs.

There are many ways to set up a machine pipe learning program to help business, and one option is to handle a cloud supply. There are many benefits of building and sending a machine learning models to the cloud, including measuring, cost efficiency, and simplified procedures compared to all the pipe.

Choosing a cloud supplier to a business, but in this article, we will investigate how we can stop the machine study machine on the Google Cloud (GCP) plate.

Let's get started.

Obvious Preparation

You must have a Google Account before continuing, as we will use GCP. Once you have created an account, Access Google Cloud console.

Once on the console, create a new project.

Then, before anything else, you need to set your payment configuration. The GCP platform requires you to register your payment information before performing many things in the platform, or with free trial account. You don't have to worry, but, as an example we will use will not take your free debt.

Please enter all payment information required to start the project. You can also need your tax information and credit card to ensure that they are correct.

All of us everywhere, let's start building our machine pipeline with GCP.

Obvious Google Platform Study

Creating our machine study pipeline, we will need an example. We will use the Dataset for the Heart Attack Predent from Kagle of this lesson. Download the information and keep it somewhere now.

Next, we must set our data storage, which is a learning machine to use. To do that, we should build a bucket of our data storage. Search 'cloud storage' to create a bucket. You must have a different world name. Currently, you do not need to change any default settings; Just click the creative button.

When the bucket has been created, upload your CSV file to it. If you do this well, you will see the data data within a bucket.

Next, we will build a new table we can ask us to use the first service. Search 'Bikiyery' and click 'Add Data'. Select 'Google Cloud Storage' and select the CSV file from the bucket in the front.

Fill in the information, especially where you go project, data form (create a new dataset or select one existing), and the table name. In Schema, select 'an automatically access' and create a table.

If you have successfully created, you can ask a table to see if you can enter the dataset.

Next, Search Vertex Ai and enable all recommended APIs. When that is finished, select 'Colab Enterprise'.

Select 'Create a writing book' to create a writing book that will use our simple machine reading pipe.

If you are familiar with Google Colab, the interface that will look very similar. You can add a letter of writing to an external source if you want.

With NoteBook Ready, Connecture and running time. In the meantime, the default machine type will be sufficient as we do not need many resources.

Let's start the development of our pipe with questions about data to ask at our table at Mhikhouly. First, we need to start a Buserer client with the following code.

from google.cloud import bigquery

client = bigquery.Client()

Then, let's ask our data on the maskyury table using the following code. Change the project ID, dataset, and the designed table creation previously.

# TODO: Replace with your project ID, dataset, and table name
query = """
SELECT *
FROM `your-project-id.your_dataset.heart_attack`
LIMIT 1000
"""
query_job = client.query(query)

df = query_job.to_dataframe()

Data is now in Pandas Datafriase in our text book. Let's change our target variable ('result') and be a number label.

import pandas as pd
from sklearn.model_selection import train_test_split
from sklearn.linear_model import LogisticRegression
from sklearn.metrics import accuracy_score

df['Outcome'] = df['Outcome'].apply(lambda x: 1 if x == 'Heart Attack' else 0)

Suppose, let's prepare our assessment details training.

df = df.select_dtypes('number')

X = df.drop('Outcome', axis=1)
y = df['Outcome']

X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.3, random_state=42)

⚠️ Note: df = df.select_dtypes('number') It is used to facilitate the example by discarding all non-numbered numbers. In the true state of the world, this is a fierce action that can disregard the useful features of the section. This is done here for convenience, and generally engineering or installation in the wet it was often considered.

When data is ready, let's train model and examine its performance.

model = LogisticRegression()
model.fit(X_train, y_train)

y_pred = model.predict(X_test)
print(f"Model Accuracy: {accuracy_score(y_test, y_pred)}")

The correct accuracy is only 0.5. This certainty can certainly be developed, but because of this example, we will continue with this simple model.

Now, let's use our model to make predictions and prepare for the results.

result_df = X_test.copy()
result_df['actual'] = y_test.values
result_df['predicted'] = y_pred
result_df.reset_index(inplace=True)

Finally, we will save forecasts for our model in Bobiery new table. Note that the following code will write more than the existing table when it is already available, rather than it.

# TODO: Replace with your project ID and destination dataset/table
destination_table = "your-project-id.your_dataset.heart_attack_predictions"
job_config = bigquery.LoadJobConfig(write_disposition=bigquery.WriteDisposition.WRITE_TRUNCATE)
load_job = client.load_table_from_dataframe(result_df, destination_table, job_config=job_config)
load_job.result()

Therefore, create a simple learning wheel pipe within the Vertex Ai Notebook.

To direct this process, you can edit the writing brochure to activate. Go to the actions of your brochure and select 'Schedule'.

Choose the frequency you need is notebook to work, for example, every Tuesday or the first day of the month. This is the easiest way to ensure that the machine study pipe works as required.

That's all set up a simple pipeline for the machine in GCP. There are many other production methods, which are high-prepared for the set of a pipe, such as using Beflow (KFP) pipelines or more integrated vertex Ai Pipelines.

Obvious Store

The Google Cloud platform provides a simplified users of users to set a machine study pipe. In this article, we have learned how we can set the pipeline using different clouds as cloud storage, three, and vertex Ai. By building a pipe in the NoteBook form and it is editing to activate it, we can create a simple, functional pipe.

I hope this has helped!

Cornellius Yudha Wijaya It is a scientific science manager and the database author. While working full-time in Allianz Indonesia, she likes to share the python and data advice with social media and media writing. Cornellius writes to a variety of AI and a study machine.