Blog-Kategorie
Cloud Solutions

Hier bloggen sepago Experten über: Cloud Solutions

| |

Improved Datamining using Azure Machine Learning #AzureML – Part 2 – Technical Details

Techincal Details

The analysis of new tweets based on “learning” is built using a 4-tier application in Microsoft Azure:

Azure machine learning API:

AzureML provides the prediction model and can be accessed through the web job using the API.

Web job in Azure:

The job (c#) reads new tweets once an hour and passes them to AzureML for classification. The prediction of each tweet is stored in the Azure SQL database.

| |

Improved Datamining using Azure Machine Learning #AzureML – Part 1 – Introduction

This is the English version of my German article on Microsoft Tech Hub: https://www.microsoft.com/germany/technet/case-studies/verbesserte-informationsbeschaffung-durch-azure-machine-learning.aspx

Sepago as a technology and consulting company located in Cologne, Hamburg and Munich stands out due to its expertise in Automation of Cloud and Application Infrastructure. Sepago’s  IT infrastructure serving  their  more than 70 employees as well as external partners is provided through Microsoft Azure and additionally on premise servers, which make desktops and applications for notebooks, tablets and smartphones (Windows phone,

| |

Improved Datamining using Azure Machine Learning #AzureML – Part 3 – Build the prediction model

Build the prediction model

Using “Setup Web-services / Predictive Web Service (Recommended)” AzureML creates the Predictive Experiment based on the training model.

Web Service Input” and “Web Service Output” are added to the predictive experiment as well in order to display input (Tweets) and output (Forecast).

Between “Sore Model” and “Web Service Output” the function “Project Columns” is inserted and configured.

| |

Fun with Microsoft azure machine learning #AzureML – scenario: weather prediction

In December I gave a rather entertaining lecture on Azure Machine Learning on the sepagoForum in Cologne: „Fun with Azure Machine Learning“. In addition to a brief introduction to neural networks and AzureML, I presented different scenarios. In these scenarios I figured out, whether neural networks may or may not solve specific tasks. One scenario is presented here.

Scenario: Weather/temperature prediction for tomorrow

Neural networks are able to make predictions on the basis of what they have learned.

| |

Import-Module AzureRM or PackageManagement fails as User #PowerShell, #PowerShellGet, #PackageManagement

The use of Azure Key Vault requires the use of PowerShell (today). Prerequisite is the PowerShell module AzureRM. This can be installed by PowerShell Package Manager using the PowerShell PowerShell Gallery (see https://azure.microsoft.com/en-us/documentation/articles/powershell-install-configure). With my local admin account it’s possible to import the PowerShell PackageManagement but not with my normal user account.

As user:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15

PS H:\> Import-Module PackageManagement

import-module : The specified module ‚PackageManagement‘ was not loaded because no valid module file was found in any module directory.

| |

Import-Module AzureRM oder PackageManagement schlägt als Benutzer fehl #PowerShell, #PowerShellGet, #PackageManagement

Die Nutzung von Azure Key Vault ist aktuell nur mit PowerShell möglich. Voraussetzung ist das PowerShell Modul AzureRM. Dieses kann mit dem PowerShell Package Manager aus der PowerShell Gallery installiert werden (https://azure.microsoft.com/en-us/documentation/articles/powershell-install-configure). Mit meinem lokalen Admin Account ist dies problemlos möglich – jedoch nicht, mit meinem normalen Benutzer Account.

Als Benutzer:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15

PS H:\> Import-Module PackageManagement

Import-Module : Das angegebene Modul „PackageManagement“ wurde nicht geladen,

| |

OneDrive for Business: Error while synchronizing files on Windows 10 – (Red Cross, Office 365 synchronization problems)

OneDrive for Business synchronized files from SharePoint Online with your local file storage. So you can work with this files on- and offline. This mechanism is very robust. Nevertheless, it can happen that the synchronization stops: affected folder are marked with a red cross – synchronization problems with Office 365 are occurred.

It’s helpful to clear the OneDrive cache and rebuild it (do not delete the synchronized folder structure – this can cause in a loss of your files on SharePoint!).

| |

OneDrive for Business: Fehler beim Synchronisieren von Dateien auf Windows 10 – (rotes Kreuz, Office 365 Synchronisierungsprobleme)

OneDrive for Business synchronisiert Dateien auf SharePoint online mit der lokalen Dateiablage. So sind die Daten offline les- und schreibbar. Dieser Mechanismus ist sehr robust. Trotzdem kann es vorkommen, dass die Synchronisierung stoppt: Betroffene Ordner werden mit einem roten Kreuz versehen – Synchronisierungsprobleme mit Office 365.

Als hilfreich hat es sich erwiesen, den Cache zu löschen und neu aufbauen zu lassen (nicht die synchronisierte Ordnerstruktur löschen – das kann zu Dateiverlusten im SharePoint führen!).

| |

Fun with Microsoft Azure Machine Learning #AzureML – Szenario: Wettervorhersage

Microsoft hat meine Success Story zu Microsoft Azure Machine Learning auf dem TechNet / IT Pro Hub veröffentlicht:

Auszug aus dem Inhalt:

Zum Erkennen von IT-Trends verwendet die sepago ein auf Azure abgebildetes Analyseverfahren, welches bestimmte Tweets analysiert. Dieses Projekt „Twends“ wurde bereits auf dem IT Pro Hub in einer eigenen Case Study beschrieben. Das Projekt soll erweitert werden und interessante/wichtige Tweets aus dem Informationsstrom von Twitter direkt hervorheben und anzeigen. 

| |

Resolved: Error while using Azure Logic Apps to retrain Azure Machine Learning – Permissions for service „AzureMLLogicAppConnector“ are set to internal but this request was external

I used Azure Logic App to retrain an experiment in AzureML regularly. Therefore, I build a logic app to do this. If I start the logic app I got the following error:

Permissions for service „AzureMLLogicAppConnector“ are set to internal but this request was external

1
2
3
4
5

„body“: {
        „status“: 403,
        „source“: „https://twxressc33ceb5b9891445aba4a46366f9a9761.azurewebsites.net/api/CheckStatus?url=&api=mMSdA80GHmzISyG?????????????????????????????lUmQHx5K%2ByNANYmzx4GuyuUwIjL????????????????KOQ%3D%3D“,
        „message“: „Permissions for service \“AzureMLLogicAppConnector\“ are set to internal but this request was external.“
    }

The following resolve this error: Set Access Level to “Public (authenticated)” in application settings

Resource group / Resources / AzureMLLogicAppConnector / All settings / Application settings

I think the reason is the different location of the resource group and the endpoint.