User

ATH: 1000

About iMLOv

We believe that everyone without professional training in informatics is able to master machine learning skills as long as the user-interface is friendly enough. By integrating several well-known machine learning algorithms into an easy-to-use interface, we present the iMLOv system. Using iMLOv can be just making several "next step" clicks. Moreover, this system accepts Excel-compatible CSV files of a simple "one row, one case" format.

Machine learning is an important technique for data classification and making predictions/decisions based on known data. It is also a key to big data analysis and artificial intelligence. In this digital era when all types of data increase explosively (proteomics, genomics, public health…), it is expectable that learning machine learning will become necessary for biologists and medical scientists in the near future. However, most machine learning software is inconvenient for users unfamiliar with programming or command-line operating. Hence, we develop this non-expert-friendly system iMLOv, which integrates decision tree, support vector machine, artificial neural network, and random forest algorithms. It also utilizes a genetic algorithm to help perform feature selections. Hopefully this system will become a useful tool for scientists to move their fields forward.

Notice: this is only a testing platform. All data has been calculated already.

Welcome to experience AIgatha AI pridictive platform,
please choose a project you interested.

Dota2 Games

Have you ever imagined that the moment characters are chosen is also the moment that you know the result of the game? Collect simple data of both characters, and you can foresee the ending of the game!

Raw data : 92650

Features : 116

Downloading dateset

Bank

Have you ever imagined that the bank will know what financial commodity you will buy without asking you first? By collect enough data and analyze it, they can predict what you will do next stop!


Raw data: 4119

Features: 16

Downloading dateset

CPBL Games

In previous game prediction, people regarded a team or an individual’s factor as indispensable data. But have you ever imagined that you can foresee the result of CPBL Games by only collecting environmental parameters of the games?

Raw data: 330

Features: 15

Downloading dateset

Setting of the Task(Dota)


Information of your training data

Training data: 92650  Prediction: 10294

Feature: 116


Testing


Groping method :

Setting of the Task(Bank)


Information of your training data

Training data: 4119  Prediction: 41188

Feature: 16


Testing


Groping method :

Setting of the Task(CPBL)


Information of your training data

Training data: 330  Prediction: 30

Feature: 15


Testing


Groping method :

Job ID (Dota):

0x0Ec8Baafa891

Return time:

2 hr

Job size:

25 Mb

Job ID (Bank):

0x1543d0F83489

Return time :

2 hr

Job size:

20 Mb

Job ID (CPBL):

0xca7AEB423e79

Return time :

2 hr

Job size:

3 Mb

Data

Training filedota2_train.csv

Prediction filedota2_prediction_test.csv

Case filesNull

Model fileNull

Tasks to perform

Features filesTraining

MethodN-fold cross validation

Number of fold3

Grouping methodInterval

Algorithms to integrate

Your algorithms list

Support Vector Machine

Artificial Neural Network

Decision Tree

Setting of the selected algorithms

Support Vector Machine

Ecological groups10

Balance modelmax*1

OptimizationYes

Log2g trial valuesiMLOv default

Artificial Neural Network

Ecological groups20

Balance modelmax*1

OptimizationYes

Decision Tree

Ecological groups100

Balance modelmax*1

OptimizationYes

Bank

Training filebank_train.csv

Prediction filebank_prediction_test.csv

Case filesNull

Model fileNull

Tasks to perform

Features filesTraining

MethodN-fold cross validation

Number of fold3

Grouping methodInterval

Algorithms to integrate

Your algorithms list

Support Vector Machine

Artificial Neural Network

Decision Tree

Setting of the selected algorithms

Support Vector Machine

Ecological groups10

Balance modelmax*1

OptimizationYes

Log2g trial valuesiMLOv default

Artificial Neural Network

Ecological groups20

Balance modelmax*1

OptimizationYes

Decision Tree

Ecological groups100

Balance modelmax*1

OptimizationYes

CPBL

Training filecpbl_train.csv

Prediction filecpbl_prediction_test.csv

Case filesNull

Model fileNull

Tasks to perform

Features filesTraining

MethodN-fold cross validation

Number of fold3

Grouping methodInterval

Algorithms to integrate

Your algorithms list

Support Vector Machine

Artificial Neural Network

Decision Tree

Setting of the selected algorithms

Support Vector Machine

Ecological groups10

Balance modelmax*1

OptimizationYes

Log2g trial valuesiMLOv default

Artificial Neural Network

Ecological groups20

Balance modelmax*1

OptimizationYes

Decision Tree

Ecological groups100

Balance modelmax*1

OptimizationYes

Dear Guest, welcome back! The queried job (0x0Ec8Baafa891) has been finshed at

The pie chart shows the evaluations of each machine learning algorithm. You can downdoad files for predicitng information.

Integrated

Prediction: 10294

Correct rate: 60.11%

Error rate: 39.89%

Downloading prediction

AUC

0.639

SEN

0.554

SPC

0.639

FPR

0.361

FNR

0.446

MCC

0.194

Support Vector Machine

Prediction: 10294

Correct rate: 59.42%

Error rate: 40.58%

Downloading prediction

AUC

0.630

SEN

0.510

SPC

0.670

FPR

0.330

FNR

0.490

MCC

0.182

Artificial Neural Network

Prediction: 10294

Correct rate: 56.94%

Error rate: 40.36%

Downloading prediction

AUC

0.608

SEN

0.633

SPC

0.669

FPR

0.329

FNR

0.481

MCC

0.189

Decision Tree

Prediction: 10294

Correct rate: 58.38%

Error rate: 41.62%

Downloading prediction

AUC

0.622

SEN

0.581

SPC

0.596

FPR

0.414

FNR

0.409

MCC

0.177

Dear Guest, welcome back! The queried job (0x1543d0F83489) has been finshed at

The pie chart shows the evaluations of each machine learning algorithm. You can downdoad files for predicitng information.

INTEGRATED

Prediction: 41188

Correct rate: 91.74%

Error rate: 8.26%

Downloading prediction

AUC

0.943

SEN

0.972

SPC

0.491

FPR

0.509

FNR

0.028

MCC

0.538

Support Vector Machine

Prediction: 41188

Correct rate: 90.82%

Error rate: 9.18%

Downloading prediction

AUC

0.985

SEN

0.985

SPC

0.785

FPR

0.225

FNR

0.050

MCC

0.685

Artificial Neural Network

Prediction: 41188

Correct rate: 90.5%

Error rate: 9.50%

Downloading prediction

AUC

0.922

SEN

0.958

SPC

0.488

FPR

0.512

FNR

0.042

MCC

0.486

Decision Tree

Prediction: 41188

Correct rate: 91.52%

Error rate: 8.48%

Downloading prediction

AUC

0.934

SEN

0.945

SPC

0.684

FPR

0.316

FNR

0.055

MCC

0.599

Dear Guest, welcome back! The queried job (0xca7AEB423e79) has been finshed at

The pie chart shows the evaluations of each machine learning algorithm. You can downdoad files for predicitng information.

INTERGRATRESD

Prediction: 40

Correct rate: 90.00%

Error rate: 10.00%

Downloading prediction

AUC

0.877

SEN

0.905

SPC

0.842

FPR

0.158

FNR

0.095

MCC

0.750

Support Vector Machine

Prediction: 40

Correct rate: 80.00%

Error rate: 20.00%

Downloading prediction

AUC

0.822

SEN

0.905

SPC

0.737

FPR

0.263

FNR

0.095

MCC

0.654

Artificial Neural Network

Prediction: 40

Correct rate: 77.50%

Error rate: 22.50%

Downloading prediction

AUC

0.905

SEN

0.762

SPC

0.842

FPR

0.158

FNR

0.238

MCC

0.604

Decision Tree

Prediction: 40

Correct rate: 85.00%

Error rate: 15.00%

Downloading prediction

AUC

0.860

SEN

0.905

SPC

0.737

FPR

0.263

FNR

0.095

MCC

0.654

Thanks for your testing. We hope you more realize how AIgatha will popular AI in daily life throught the process.
In the future, data competition, rather than algorithm competion, will be the maintream.
We believe that what is valuable is not the tool itself. It is using the tool to resolve business problmes that really creates values. We hope everyone can find out their own proper AI algorithm and eigenvalues with our platform. And we hope more start-up companies can concentrate on developing algorithms and sell them on our platform.

In the field of machine learning, We establish AI testing platform with Training ,examine the effectiveness after training through Testing, and test if the model is flexible enough for unknown data in the future through .Independent testing. Finally, we get get the answers through Prediction.

Choose the machine learning algorithm. This will affect calculating fee and accuracy. We suggest that you operate all selections at first time to realize the proper algorithm.

Upload dada resource. We have offered demo on the homepage. Users can take CSV file in zip file as reference.

Finishied

Task: Choose how many Project you want to divide.
Please note that the amount of Project is not directly proportional to calculating time.
Miner: Choose how many miners are needed to verify a task.
Each task needs at least two miners for verification

Algorithms




Optimization:



Optimization:



Optimization:





Chsose the task

Some functions are limited because this platform is only for display.

Project will be uploaded to the nearest Chronos Fortress by users’ computer and will be calculated, verified and defaulted.

Chronos Fortress will divide the task and report the calculating fee of Project to users.

2 hr
70 ATH