Training artificial neural networks directly on the concordance index for censored data using genetic algorithms

Kalderstam, Jonas; Edén, Patrik; Bendahl, Pär‐Ola; Strand, Carina; Fernö, Mårten; Ohlsson, Mattias

doi:10.1016/j.artmed.2013.03.001

Cited by 25 publications

(27 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Two different computational models were tested to investigate the effect of regional BSI measurements. The standard Cox proportional hazards model and one based on artificial neural networks optimizing directly on the C-index performance measure [11]. The C-index was used in this study to evaluate regional BSI values since it directly measures the models’ ability to sort the patients according to survival time, which can be useful for subsequent risk group definitions.…”

Section: Discussionmentioning

confidence: 99%

“…To investigate this possibility we turned to a model based on artificial neural networks (ANN) introduced in [11]. ANN is a machine learning technique that offers strong non-linear capabilities.…”

Section: Methodsmentioning

confidence: 99%

“…We used Cox proportional hazards in the survival modeling to investigate the effect of going from the total BSI measurement to twelve regional ones. To allow for possible non-linear interactions between regional BSI measurements we also employed a more flexible modeling approach based on artificial neural networks [11]. …”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Analysis of regional bone scan index measurements for the survival of patients with prostate cancer

et al. 2014

Self Cite

View full text Add to dashboard Cite

BackgroundA bone scan is a common method for monitoring bone metastases in patients with advanced prostate cancer. The Bone Scan Index (BSI) measures the tumor burden on the skeleton, expressed as a percentage of the total skeletal mass. Previous studies have shown that BSI is associated with survival of prostate cancer patients. The objective in this study was to investigate to what extent regional BSI measurements, as obtained by an automated method, can improve the survival analysis for advanced prostate cancer.MethodsThe automated method for analyzing bone scan images computed BSI values for twelve skeletal regions, in a study population consisting of 1013 patients diagnosed with prostate cancer. In the survival analysis we used the standard Cox proportional hazards model and a more advanced non-linear method based on artificial neural networks. The concordance index (C-index) was used to measure the performance of the models.ResultsA Cox model with age and total BSI obtained a C-index of 70.4%. The best Cox model with regional measurements from Costae, Pelvis, Scapula and the Spine, together with age, got a similar C-index (70.5%). The overall best single skeletal localisation, as measured by the C-index, was Costae. The non-linear model performed equally well as the Cox model, ruling out any significant non-linear interactions among the regional BSI measurements.ConclusionThe present study showed that the localisation of bone metastases obtained from the bone scans in prostate cancer patients does not improve the performance of the survival models compared to models using the total BSI. However a ranking procedure indicated that some regions are more important than others.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Methodsmentioning

confidence: 99%

See 1 more Smart Citation

Analysis of regional bone scan index measurements for the survival of patients with prostate cancer

et al. 2014

Self Cite

View full text Add to dashboard Cite

show abstract

“…The United State Environmental Agency suggested substitution if censored data is less than 15% of the total data-set (63,64). [8], BDL, for example x < 1.1, were multiplied by the factor 0.75 to give 0.825.…”

Section: Censored Datamentioning

confidence: 99%

Validity and Errors in Water Quality Data — A Review

Rangeti¹,

Dzwairo²,

Barratt³

et al. 2015

Research and Practices in Water Quality

View full text Add to dashboard Cite

“…Note that any activation function can be used instead, including y = x, but derivatives are required if any gradiant based algorithm performed for the optimization. Evolutionary algorithms, such as genetic algorithms (Goldberg & Holland, 1988) and differential evolution (Storn & Price, 1997) can also be applied to estimate unknown parameters w i (Kalderstam et al, 2013;Ilonen et al, 2003).…”

Section: Suppose the Regression Model Ismentioning

confidence: 99%

Selecting Location of Retail Stores Using Artificial Neural Networks and Google Places API

Satman

Altunbey

2014

IJSP

View full text Add to dashboard Cite

Selection of locations of retail stores (similarly, facilities) is a kind of decision problem in which the researcher targets maximum profit or minimum cost by using statistical or judgmental forecasts. In many cases, statistical forecasts are based on survey data or data collected for an other store which are not reliable as the survey may not measure the potential demand or the other store that is used for comparison may not be similar enough. Addition to this, estimating the potential demand for a non-existing store can be expensive. Google Places API is a software library which provides a nearby search utility that is able to report several kinds of institutions, establishments and other objects placed in a circumferential area for a given radius. In this paper, we suggest a method which is based on estimating the relation between environmental properties of existing retail stores and their financial rankings. We remark that feed forward neural networks are capable to reveal such a relation as they permit using both integer, real and binary variables in the model and non-linearity. We apply our method on the location selection problem of a retail stores chain in Turkey (Note 1) which has 144 clients in its organization. Expert opinions in the organization show us applicability of this method is high in similar situations.

show abstract

Training artificial neural networks directly on the concordance index for censored data using genetic algorithms

Cited by 25 publications

References 23 publications

Analysis of regional bone scan index measurements for the survival of patients with prostate cancer

Analysis of regional bone scan index measurements for the survival of patients with prostate cancer

Validity and Errors in Water Quality Data — A Review

Selecting Location of Retail Stores Using Artificial Neural Networks and Google Places API

Contact Info

Product

Resources

About