ISLAB/CAISR - User contributions [en]

Meta-learning for evaluation and implementation of predictive maintenance solution

2018-10-09T09:13:37Z

Pabmor:

{{StudentProjectTemplate
|Summary=Finding the best strategy to schedule a predictive maintenance intervention optimizing the cost of unexpected breakdowns against unuseful visits to the workshop.
|Keywords=Predictive maintenance, metalearning, classification, regression.
|TimeFrame=Winter 2018 / Summer 2019
|References=[1] Rune Prytz. Machine learning methods for vehicle predictive maintenance using o-board and on-board data. Licentiate thesis, Halmstad University Press, 2014.
[2] Rune Prytz, Slawomir Nowaczyk, Thorsteinn Rögnvaldsson, and Stefan Byttner. Predicting the need for vehicle compressor repairs using maintenance records and logged vehicle data. Engineering applications of articial intelligence, 41:139{150, 2015.

|Prerequisites=Good knowledge of applied data science: classification and regression. Basic knowledge of optimization.
|Supervisor=Pablo del Moral, Sławomir Nowaczyk,
|Level=Master
|Status=Open
}}
Background: there are two main approaches towards predictive maintenance.

- The first one involves classification: looking at historical data of a machine we label as faulty the data happening before a breakdown happened. We will later predict the probability of being faulty, trying to catch a breakdown before it happens.

- The second one involves regression: looking at historical data we label our data with the remaining time to failure. We will later calculate the estimated time to failure.

Usually the work of the data scientists would end here, but there are many more issues arising after our models make their prediction.

When do we send the faulty machine to the workshop? Sending a machine to the workshop or sending a technician to repair the machine has a cost in terms of money and downtime, especially if the machine is not broken.

Should we always send the machine to repair every time our models send an alarm? Should we wait until the next reading and check if the machine is still predicted to be faulty? What if the next reading is regarded as healthy by our classifier? What if the estimated time to failure is small but does not evolve with time? What if the time to failure is big, but is decreasing fast?

There is a big gap between the machine learning models that predict failures and its implementation in the industry. A framework to develop strategies for implementing these solutions and optimization of the cost of unplanned breakdowns and unnecessary repairs is needed for a successful final product.

Replacements vs repairs. What is really a breakdown from the point of view of the data?

2018-10-09T09:13:18Z

Pabmor:

{{StudentProjectTemplate
|Summary=Finding ways to test how similar or different are the actions performed on a faulty machine. How does the effect of an intervention on the machine affect the data?
|Keywords=Predictive maintenance, classification, regression, clustering.
|TimeFrame=winter 2018/summer 2019
|Prerequisites=Good knowledge of applied data science: supervised and unsupervised learning.
|Supervisor=Pablo del Moral, Sławomir Nowaczyk,
|Level=Master
|Status=Open
}}
When a machine breaks down a technician makes a diagnosis and performs some corrective action. This action can be either a repair, where no component is replaces; or a replacement is performed, where the faulty component is replaced. On top of that, the machine can also undergo preventive maintenance actions, where components are replaced even if they have not broken yet.
When building a predictive maintenance system, the data scientist usually takes the replacement of a component as a failure and tends to forget about other actions that could have been performed. The validity of this assumption will be critical in determining the success of the predictive maintenance models.
The goal of this project is to develop the techniques necessaries to test it, using a dataset coming from a fleet of sterilizers with partial service records where many different components are considered. Understanding the differences between repairs, replacements and preventive maintenance actions is a need to be successful in the industrial application.

Meta-learning for evaluation and implementation of predictive maintenance solution

2018-10-09T09:12:58Z

Pabmor:

{{StudentProjectTemplate
|Summary=Finding the best strategy to schedule a predictive maintenance intervention optimizing the cost of unexpected breakdowns against unuseful visits to the workshop.
|Keywords=Predictive maintenance, metalearning, classification, regression.
|TimeFrame=Winter 2017 / Summer 2018
|References=[1] Rune Prytz. Machine learning methods for vehicle predictive maintenance using o-board and on-board data. Licentiate thesis, Halmstad University Press, 2014.
[2] Rune Prytz, Slawomir Nowaczyk, Thorsteinn Rögnvaldsson, and Stefan Byttner. Predicting the need for vehicle compressor repairs using maintenance records and logged vehicle data. Engineering applications of articial intelligence, 41:139{150, 2015.

|Prerequisites=Good knowledge of applied data science: classification and regression. Basic knowledge of optimization.
|Supervisor=Pablo del Moral, Sławomir Nowaczyk,
|Level=Master
|Status=Open
}}
Background: there are two main approaches towards predictive maintenance.

- The first one involves classification: looking at historical data of a machine we label as faulty the data happening before a breakdown happened. We will later predict the probability of being faulty, trying to catch a breakdown before it happens.

- The second one involves regression: looking at historical data we label our data with the remaining time to failure. We will later calculate the estimated time to failure.

Usually the work of the data scientists would end here, but there are many more issues arising after our models make their prediction.

When do we send the faulty machine to the workshop? Sending a machine to the workshop or sending a technician to repair the machine has a cost in terms of money and downtime, especially if the machine is not broken.

Should we always send the machine to repair every time our models send an alarm? Should we wait until the next reading and check if the machine is still predicted to be faulty? What if the next reading is regarded as healthy by our classifier? What if the estimated time to failure is small but does not evolve with time? What if the time to failure is big, but is decreasing fast?

There is a big gap between the machine learning models that predict failures and its implementation in the industry. A framework to develop strategies for implementing these solutions and optimization of the cost of unplanned breakdowns and unnecessary repairs is needed for a successful final product.

Meta-learning for evaluation and implementation of predictive maintenance solution

2018-10-09T09:12:23Z

Pabmor:

{{StudentProjectTemplate
|Summary=Finding the best strategy to schedule a predictive maintenance intervention optimizing the cost of unexpected breakdowns against unuseful visits to the workshop.
|Keywords=Predictive maintenance, metalearning, classification, regression.
|TimeFrame=Winter 2017 / Summer 2018
|References=[1] Rune Prytz. Machine learning methods for vehicle predictive maintenance using o-board and on-board data. Licentiate thesis, Halmstad University Press, 2014.
[2] Rune Prytz, Slawomir Nowaczyk, Thorsteinn Rögnvaldsson, and Stefan Byttner. Predicting the need for vehicle compressor repairs using maintenance records and logged vehicle data. Engineering applications of articial intelligence, 41:139{150, 2015.

|Prerequisites=Good knowledge of applied data science: classification and regression. Basic knowledge of optimization.
|Supervisor=Pablo del Moral, Sławomir Nowaczyk,
|Level=Master
|Status=Open
}}
Background: there are two main approaches towards predictive maintenance.
- The first one involves classification: looking at historical data of a machine we label as faulty the data happening before a breakdown happened. We will later predict the probability of being faulty, trying to catch a breakdown before it happens.
- The second one involves regression: looking at historical data we label our data with the remaining time to failure. We will later calculate the estimated time to failure.
Usually the work of the data scientists would end here, but there are many more issues arising after our models make their prediction.
When do we send the faulty machine to the workshop? Sending a machine to the workshop or sending a technician to repair the machine has a cost in terms of money and downtime, especially if the machine is not broken.
Should we always send the machine to repair every time our models send an alarm? Should we wait until the next reading and check if the machine is still predicted to be faulty? What if the next reading is regarded as healthy by our classifier? What if the estimated time to failure is small but does not evolve with time? What if the time to failure is big, but is decreasing fast?
There is a big gap between the machine learning models that predict failures and its implementation in the industry. A framework to develop strategies for implementing these solutions and optimization of the cost of unplanned breakdowns and unnecessary repairs is needed for a successful final product.

PRIME

2017-11-16T16:28:41Z

Pabmor:

{{ResearchProjInfo
|Title=PRIME
|ContactInformation=Pablo del Moral
|ShortDescription=PRedictive Intelligent Maintenance Enabler
|Description=The goal of this project is predicting failures in a fleet of sterilizers deployed in hospitals all over the world. The characteristics of this problem are general to the field of predictive maintenance for different application fields.

Companies are interested in predictive maintenance to reduce the down time of their machines. In general the list of critical components, whose unexpected breakdowns would result in stopping the machine, is long. Therefore, the scope of a predictive maintenance system should be predicting failures in a big number of different components.

For several years, systems such as cars, sterilizers or industrial equipment have been equipped with a significant amount of sensors. Which signals to record is in general not decided based on the predictive maintenance needs, but on the requirements of security or controllers among other reasons. The sensors mounted usually don’t describe the particular behavior of the components of interest, but measure physical quantities that can be influenced by the different behavior of several components.

Predicting what component will fail when requires historic data about the operation of the machines, but also needs to be linked to the occurrence of failures, so that we can label the recorded data. In general, companies have access and store data coming from their machines, but don’t necessary have access to the whole history of repairs. The owner of the machines can decide whether to perform maintenance and repairs with the official service or any other unofficial service.

The main research goal of this project is to build a framework that allows predicting all type of failures that can happen in a machine.
|LogotypeFile=Sterilizer photo.png
|ProjectResponsible=Pablo
|ProjectStart=2016/09/01
|ProjectEnd=2021/03/01
|ApplicationArea=Intelligent Vehicles
}}
{{AssignProjPartner
|projectpartner=Test
}}
__NOTOC__
{{ShowResearchProject}}

[[Image:Sterilizer photo.png|150px]]

PRIME

2017-11-16T16:26:37Z

Pabmor:

{{ResearchProjInfo
|Title=PRIME
|ContactInformation=Pablo del Moral
|ShortDescription=PRedictive Intelligent Maintenance Enabler
|Description=The goal of this project is predicting failures in a fleet of sterilizers deployed in hospitals all over the world. The characteristics of this problem are general to the field of predictive maintenance for different application fields.

Companies are interested in predictive maintenance to reduce the down time of their machines. In general the list of critical components, whose unexpected breakdowns would result in stopping the machine, is long. Therefore, the scope of a predictive maintenance system should be predicting failures in a big number of different components.

For several years, systems such as cars, sterilizers or industrial equipment have been equipped with a significant amount of sensors. Which signals to record is in general not decided based on the predictive maintenance needs, but on the requirements of security or controllers among other reasons. The sensors mounted usually don’t describe the particular behavior of the components of interest, but measure physical quantities that can be influenced by the different behavior of several components.

Predicting what component will fail when requires historic data about the operation of the machines, but also needs to be linked to the occurrence of failures, so that we can label the recorded data. In general, companies have access and store data coming from their machines, but don’t necessary have access to the whole history of repairs. The owner of the machines can decide whether to perform maintenance and repairs with the official service or any other unofficial service.

The main research goal of this project is to build a framework that allows predicting all type of failures that can happen in a machine.
|LogotypeFile=Sterilizer photo.png
|ProjectResponsible=Pablo
|ProjectStart=2016/09/01
|ProjectEnd=2021/03/01
|ApplicationArea=Intelligent Vehicles
}}
__NOTOC__
{{ShowResearchProject}}

[[Image:Sterilizer photo.png|150px]]

PRIME

2017-11-16T16:26:03Z

Pabmor:

{{ResearchProjInfo
|Title=PRIME
|ContactInformation=Pablo del Moral
|ShortDescription=PRedictive Intelligent Maintenance Enabler
|LogotypeFile=Sterilizer photo.png
|ProjectResponsible=Pablo
|ProjectStart=2016/09/01
|ProjectEnd=2021/03/01
|ApplicationArea=Intelligent Vehicles
}}
__NOTOC__
{{ShowResearchProject}}

[[Image:Sterilizer photo.png|150px]]

The goal of this project is predicting failures in a fleet of sterilizers deployed in hospitals all over the world. The characteristics of this problem are general to the field of predictive maintenance for different application fields.

Companies are interested in predictive maintenance to reduce the down time of their machines. In general the list of critical components, whose unexpected breakdowns would result in stopping the machine, is long. Therefore, the scope of a predictive maintenance system should be predicting failures in a big number of different components.

For several years, systems such as cars, sterilizers or industrial equipment have been equipped with a significant amount of sensors. Which signals to record is in general not decided based on the predictive maintenance needs, but on the requirements of security or controllers among other reasons. The sensors mounted usually don’t describe the particular behavior of the components of interest, but measure physical quantities that can be influenced by the different behavior of several components.

Predicting what component will fail when requires historic data about the operation of the machines, but also needs to be linked to the occurrence of failures, so that we can label the recorded data. In general, companies have access and store data coming from their machines, but don’t necessary have access to the whole history of repairs. The owner of the machines can decide whether to perform maintenance and repairs with the official service or any other unofficial service.

The main research goal of this project is to build a framework that allows predicting all type of failures that can happen in a machine.

PRIME

2017-11-16T16:25:00Z

Pabmor:

{{ResearchProjInfo
|Title=PRIME
|ContactInformation=Pablo del Moral
|ShortDescription=PRedictive Intelligent Maintenance Enabler
|Description=The goal of this project is predicting failures in a fleet of sterilizers deployed in hospitals all over the world. The characteristics of this problem are general to the field of predictive maintenance for different application fields.

Companies are interested in predictive maintenance to reduce the down time of their machines. In general the list of critical components, whose unexpected breakdowns would result in stopping the machine, is long. Therefore, the scope of a predictive maintenance system should be predicting failures in a big number of different components.

For several years, systems such as cars, sterilizers or industrial equipment have been equipped with a significant amount of sensors. Which signals to record is in general not decided based on the predictive maintenance needs, but on the requirements of security or controllers among other reasons. The sensors mounted usually don’t describe the particular behavior of the components of interest, but measure physical quantities that can be influenced by the different behavior of several components.

Predicting what component will fail when requires historic data about the operation of the machines, but also needs to be linked to the occurrence of failures, so that we can label the recorded data. In general, companies have access and store data coming from their machines, but don’t necessary have access to the whole history of repairs. The owner of the machines can decide whether to perform maintenance and repairs with the official service or any other unofficial service.

The main research goal of this project is to build a framework that allows predicting all type of failures that can happen in a machine.
|LogotypeFile=Sterilizer photo.png
|ProjectResponsible=Pablo
|ProjectStart=2016/09/01
|ProjectEnd=2021/03/01
|ApplicationArea=Intelligent Vehicles
}}
__NOTOC__
{{ShowResearchProject}}

[[Image:Sterilizer photo.png|150px]]

PRIME

2017-11-16T16:20:08Z

Pabmor:

{{ResearchProjInfo
|Title=PRIME
|ContactInformation=pablo.del_moral@hh.se
|ShortDescription=PRedictive Intelligent Maintenance Enabler
|Description=The goal of this project is predicting failures in a fleet of sterilizers deployed in hospitals all over the world. The characteristics of this problem are general to the field of predictive maintenance for different application fields.

Companies are interested in predictive maintenance to reduce the down time of their machines. In general the list of critical components, whose unexpected breakdowns would result in stopping the machine, is long. Therefore, the scope of a predictive maintenance system should be predicting failures in a big number of different components.

For several years, systems such as cars, sterilizers or industrial equipment have been equipped with a significant amount of sensors. Which signals to record is in general not decided based on the predictive maintenance needs, but on the requirements of security or controllers among other reasons. The sensors mounted usually don’t describe the particular behavior of the components of interest, but measure physical quantities that can be influenced by the different behavior of several components.

Predicting what component will fail when requires historic data about the operation of the machines, but also needs to be linked to the occurrence of failures, so that we can label the recorded data. In general, companies have access and store data coming from their machines, but don’t necessary have access to the whole history of repairs. The owner of the machines can decide whether to perform maintenance and repairs with the official service or any other unofficial service.

The main research goal of this project is to build a framework that allows predicting all type of failures that can happen in a machine.
|LogotypeFile=Sterilizer photo.png
|ProjectResponsible=Pablo
|ProjectStart=2016/09/01
|ProjectEnd=2021/03/01
|ApplicationArea=Predictive Maintenance
}}
__NOTOC__
{{ShowResearchProject}}

[[Image:Sterilizer photo.png|150px]]

PRIME

2017-11-16T16:19:53Z

Pabmor:

PRIME

2017-11-16T16:19:38Z

Pabmor:

PRIME

2017-11-16T16:19:16Z

Pabmor:

PRIME

2017-11-16T16:05:23Z

Pabmor:

{{ResearchProjInfo
|Title=PRIME
|ContactInformation=pablo.del_moral@hh.se
|ShortDescription=PRedictive Intelligent Maintenance Enabler
|LogotypeFile=Sterilizer photo.png
|ProjectResponsible=Pablo
|ProjectStart=2016/09/01
|ProjectEnd=2021/03/01
|ApplicationArea=Predictive Maintenance
}}
__NOTOC__
{{ShowResearchProject}}

[[Image:Sterilizer photo.png|150px]]


[[Image:Sterilizer photo.png|300px]]

PRIME

2017-11-16T16:04:36Z

Pabmor:

{{ResearchProjInfo
|Title=PRIME
|ContactInformation=pablo.del_moral@hh.se
|ShortDescription=PRedictive Intelligent Maintenance Enabler
|LogotypeFile=Sterilizer photo.png
|ProjectResponsible=Pablo
|ProjectStart=2016/09/01
|ProjectEnd=2021/03/01
|ApplicationArea=Predictive Maintenance
}}
__NOTOC__
{{ShowResearchProject}}

Swedish Research Council, Research Project No: 2016-03497

[[Image:VRlogo.png|150px]]


[[Image:Sterilizer photo.png|300px]]

PRIME

2017-11-16T16:03:01Z

Pabmor:

{{ResearchProjInfo
|Title=PRIME
|ContactInformation=pablo.del_moral@hh.se
|ShortDescription=PRedictive Intelligent Maintenance Enabler
|LogotypeFile=Sterilizer photo.png
|ProjectResponsible=Pablo
|ProjectStart=2017/01/01
|ProjectEnd=2020/12/30
|ApplicationArea=Biometrics
}}
__NOTOC__
{{ShowResearchProject}}

Swedish Research Council, Research Project No: 2016-03497

[[Image:VRlogo.png|150px]]


[[Image:Eye mobile.png|300px]]

File:Sterilizer photo.png

2017-11-16T16:02:54Z

Pabmor:

Wisdom of the crowd, is there a single crowd?

2017-09-27T19:45:26Z

Pabmor: Created page with "{{StudentProjectTemplate |Summary=Finding different clusters of machines based on their behavior as a first step towards anomaly detection and predictive maintenance. |Keyword..."

{{StudentProjectTemplate
|Summary=Finding different clusters of machines based on their behavior as a first step towards anomaly detection and predictive maintenance.
|Keywords=Predictive maintenance, anomaly detection, clustering.
|Prerequisites=Good knowledge of applied data science: supervised and unsupervised learning.
|Supervisor=Pablo del Moral, Sławomir Nowaczyk,
|Level=Master
|Status=Open
}}
Many anomaly detection methods rely on the so-called wisdom of the crowd: the behavior of the majority is regarded as normal and everything that is not similar to the majority is regarded as an anomaly.
In this project you will be working with data coming from a fleet of sterilizers used in hospitals for sterilization of medical equipment. A priori, these machines can have very different configurations and can be used for different purposes, loads, and in different conditions.
For any machine learning related task a question has to be answered, how similar are the data coming from different machines?, can we use the data from one machine to predict on another? The data coming from different machines will be different up to some degree, but it is likely that there are groups of machines behaving similarly. The next question is, how different are these groups of machines from each other?
The goal for this project is to answer those questions for a real case problem coming from one of our industrial partners.

Replacements vs repairs. What is really a breakdown from the point of view of the data?

2017-09-27T19:43:46Z

Pabmor: Created page with "{{StudentProjectTemplate |Summary=Finding ways to test how similar or different are the actions performed on a faulty machine. How does the effect of an intervention on the ma..."

{{StudentProjectTemplate
|Summary=Finding ways to test how similar or different are the actions performed on a faulty machine. How does the effect of an intervention on the machine affect the data?
|Keywords=Predictive maintenance, classification, regression, clustering.
|TimeFrame=winter 2017/summer 2018
|Prerequisites=Good knowledge of applied data science: supervised and unsupervised learning.
|Supervisor=Pablo del Moral, Sławomir Nowaczyk,
|Level=Master
|Status=Open
}}
When a machine breaks down a technician makes a diagnosis and performs some corrective action. This action can be either a repair, where no component is replaces; or a replacement is performed, where the faulty component is replaced. On top of that, the machine can also undergo preventive maintenance actions, where components are replaced even if they have not broken yet.
When building a predictive maintenance system, the data scientist usually takes the replacement of a component as a failure and tends to forget about other actions that could have been performed. The validity of this assumption will be critical in determining the success of the predictive maintenance models.
The goal of this project is to develop the techniques necessaries to test it, using a dataset coming from a fleet of sterilizers with partial service records where many different components are considered. Understanding the differences between repairs, replacements and preventive maintenance actions is a need to be successful in the industrial application.

Meta-learning for evaluation and implementation of predictive maintenance solution

2017-09-27T13:03:13Z

Pabmor: Created page with "{{StudentProjectTemplate |Summary=Finding the best strategy to schedule a predictive maintenance intervention optimizing the cost of unexpected breakdowns against unuseful vis..."

{{StudentProjectTemplate
|Summary=Finding the best strategy to schedule a predictive maintenance intervention optimizing the cost of unexpected breakdowns against unuseful visits to the workshop.
|Keywords=Predictive maintenance, metalearning, classification, regression.
|TimeFrame=Winter 2017 / Summer 2018
|References=[1] Rune Prytz. Machine learning methods for vehicle predictive maintenance using o-board and on-board data. Licentiate thesis, Halmstad University Press, 2014.
[2] Rune Prytz, Slawomir Nowaczyk, Thorsteinn Rögnvaldsson, and Stefan Byttner. Predicting the need for vehicle compressor repairs using maintenance records and logged vehicle data. Engineering applications of articial intelligence, 41:139{150, 2015.

|Prerequisites=Good knowledge of applied data science: classification and regression. Basic knowledge of optimization.
|Supervisor=Pablo del Moral, Sławomir Nowaczyk,
|Level=Master
|Status=Open
}}
Background: there are two main approaches towards predictive maintenance.
- The first one involves classification: looking at historical data of a machine we label as faulty the data happening before a breakdown happened. We will later predict the probability of being faulty, trying to catch a breakdown before it happens.
- The second one involves regression: looking at historical data we label our data with the remaining time to failure. We will later calculate the estimated time to failure.
Usually the work of the data scientists would end here, but there are many more issues arising after our models make their prediction.
When do we send the faulty machine to the workshop? Sending a machine to the workshop or sending a technician to repair the machine has a cost in terms of money and downtime, especially if the machine is not broken.
Should we always send the machine to repair every time our models send an alarm? Should we wait until the next reading and check if the machine is still predicted to be faulty? What if the next reading is regarded as healthy by our classifier? What if the estimated time to failure is small but does not evolve with time? What if the time to failure is big, but is decreasing fast?
There is a big gap between the machine learning models that predict failures and its implementation in the industry. A framework to develop strategies for implementing these solutions and optimization of the cost of unplanned breakdowns and unnecessary repairs is needed for a successful final product.

Detecting different types of machines based on usage

2016-10-26T09:17:29Z

Pabmor:

{{StudentProjectTemplate
|Summary=This project is about studying how can we distinguish among different types of machines based on their usage.
|Keywords=Data Mining. Data representation
|References=1.- Bengio Y, Courville A, P Vincent P. Representation Learning: A Review and New Perspectives. IEEE Transactions on Pattern Analysis and Machine Intelligence. Volume: 35, Issue: 8, Aug. 2013.

2.- Kotsiantis S. Supervised Machine Learning: A Review of Classification Techniques. Informatica 31 (2007) 249-268

3.- Grira N, Crucianu M, Boujemaa N. Unsupervised and Semi-supervised Clustering: a Brief Survey.

4.- Taskar B, Segal E, Koller D. Probabilistic Classification and Clustering in Relational Data.
|Prerequisites=Good knowledge of machine learning and programming skills for implementing machine learning algorithms
|Supervisor=Sławomir Nowaczyk, Pablo del Moral.
|Level=Master
|Status=Open
}}
In PRIME project we are analyzing data collected from 3016 machines sold by Getinge Group. These machines consist of sterilizers, washers (and possibly something else) running under different control units.

The data available is a time-stamped log of the programs run in every machine and the sequence of phases undergone through these programs; along with other information.

The goal for this project is finding the right data representation to apply supervised and unsupervised machine learning methods in order to complete these tasks based on the usage of the machines:

# Separate washers and sterilizers based on their usage.
# Separate different types of control units. There are basically two different control units running on the machines.
# Separate different models of machines. There are different types and models of machines in the cases of both sterilizers and washers.
# Separate different usages of the machines. Similar machines can be operating under different conditions.

Detecting different types of machines based on usage

2016-10-26T09:15:04Z

Pabmor:

{{StudentProjectTemplate
|Summary=This project is about studying how can we distinguish among different types of machines
|Keywords=Data Mining. Data representation
|References=1.- Bengio Y, Courville A, P Vincent P. Representation Learning: A Review and New Perspectives. IEEE Transactions on Pattern Analysis and Machine Intelligence. Volume: 35, Issue: 8, Aug. 2013.

2.- Kotsiantis S. Supervised Machine Learning: A Review of Classification Techniques. Informatica 31 (2007) 249-268

3.- Grira N, Crucianu M, Boujemaa N. Unsupervised and Semi-supervised Clustering: a Brief Survey.

4.- Taskar B, Segal E, Koller D. Probabilistic Classification and Clustering in Relational Data.
|Prerequisites=Good knowledge of machine learning and programming skills for implementing machine learning algorithms
|Supervisor=Sławomir Nowaczyk, Pablo del Moral.
|Level=Master
|Status=Open
}}
In PRIME project we are analyzing data collected from 3016 machines sold by Getinge Group. These machines consist of sterilizers, washers (and possibly something else) running under different control units.

The data available is a time-stamped log of the programs run in every machine and the sequence of phases undergone through these programs; along with other information.

The goal for this project is finding the right data representation to apply supervised and unsupervised machine learning methods in order to complete these tasks based on the usage of the machines:

# Separate washers and sterilizers based on their usage.
# Separate different types of control units. There are basically two different control units running on the machines.
# Separate different models of machines. There are different types and models of machines in the cases of both sterilizers and washers.
# Separate different usages of the machines. Similar machines can be operating under different conditions.

Detecting different types of machines based on usage

2016-10-25T16:46:44Z

Pabmor:

{{StudentProjectTemplate
|Summary=Detecting different types of machines based on usage
|Keywords=Data Mining. Data representation
|References=1.- Wang X, Mueen A, Keogh E, Ding H, Trajcevski G, Scheuerman P. Experimental comparison of representation methods and distance measures for time series data. Data Min Knowl Disc (2013) 26:275–309

2.- Kotsiantis S. Supervised Machine Learning: A Review of Classification Techniques. Informatica 31 (2007) 249-268

3.- Grira N, Crucianu M, Boujemaa N. Experimental comparison of representation methods and distance measures for time series data. Data Min Knowl Disc (2013) 26:275–309

4.- Taskar B, Segal E, Koller D. Probabilistic Classification and Clustering in Relational Data.
|Supervisor=Sławomir Nowaczyk, Pablo del Moral,
|Level=Master
|Status=Open
}}
In PRIME project we are analyzing data collected from 3016 machines sold by Getinge Group. These machines consist of sterilizers, washers (and possibly something else) running under different control units.

The data available is a time-stamped log of the programs run in every machine and the sequence of phases undergone through these programs; along with other information.

The goal for this project is finding the right data representation to apply supervised and unsupervised machine learning methods in order to complete these tasks based on the usage of the machines:

- separate different types of machines.
- separate different types of control units.
- separate different models of machines.
- ¿¿¿ separate different usages of the machines??

Detecting different types of machines based on usage

2016-10-25T16:36:44Z

Pabmor:

{{StudentProjectTemplate
|Summary=Detecting different types of machines based on usage
|Keywords=Data Mining. Data representation
|References=1.- Wang X, Mueen A, Keogh E, Ding H, Trajcevski G, Scheuerman P. Experimental comparison of representation methods and distance measures for time series data. Data Min Knowl Disc (2013) 26:275–309

2.- Kotsiantis S. Supervised Machine Learning: A Review of Classification Techniques. Informatica 31 (2007) 249-268

3.- Grira N, Crucianu M, Boujemaa N. Experimental comparison of representation methods and distance measures for time series data. Data Min Knowl Disc (2013) 26:275–309

4.- Taskar B, Segal E, Koller D. Probabilistic Classification and Clustering in Relational Data.
|Supervisor=Sławomir Nowaczyk, Pablo del Moral,
}}
In PRIME project we are analyzing data collected from 3016 machines sold by Getinge Group. These machines consist of sterilizers, washers (and possibly something else) running under different control units.

The data available is a time-stamped log of the programs run in every machine and the sequence of phases undergone through these programs; along with other information.

The goal for this project is finding the right data representation to apply supervised and unsupervised machine learning methods in order to complete these tasks based on the usage of the machines:

- separate different types of machines.
- separate different types of control units.
- separate different models of machines.
- ¿¿¿ separate different usages of the machines??

Pablo del Moral

2016-09-22T10:41:04Z

Pabmor:

{{Person
|Family Name=Del Moral Pastor
|Given Name=Pablo José
|Title=M. Sc.
|Phone=+46-72-977-36-39
|Position=PhD Candidate
|Email=pablo.del_moral@hh.se
|Image=Pablo José Del Moral Pastor web.jpg
|Office=E502
}}
[[Category:Staff]]


{{ShowPerson}}
{{InsertProjects}}

Pablo del Moral

2016-09-22T08:45:01Z

Pabmor:

{{Person
|Family Name=Del Moral Pastor
|Given Name=Pablo José
|Phone=+46-72-977-36-39
|Position=PhD Candidate
|Email=pablo.del_moral@hh.se
|Image=Pablo José Del Moral Pastor web.jpg
}}
[[Category:Staff]]


{{ShowPerson}}
{{InsertProjects}}

File:Pablo José Del Moral Pastor web.jpg

2016-09-22T08:35:57Z

Pabmor:

Pablo del Moral

2016-09-19T10:21:17Z

Pabmor: Created page with "{{Person |Family Name=Del Moral Pastor |Given Name=Pablo José |Phone=+46-72-977-36-39 |Position=PhD Candidate |Email=pablo.del_moral@hh.se }} Category:Staff <!--Remove o..."

{{Person
|Family Name=Del Moral Pastor
|Given Name=Pablo José
|Phone=+46-72-977-36-39
|Position=PhD Candidate
|Email=pablo.del_moral@hh.se
}}
[[Category:Staff]]


{{ShowPerson}}
{{InsertProjects}}