![Institutional Repository [SANDBOX]](images/dias.logo.png "Institutional Repository [SANDBOX]"){kind=logo}
Institutional Repository [SANDBOX]
Technical University of Crete
Technical University of Crete
EN
|
EL
| URI: | http://purl.tuc.gr/dl/dias/4B7EC6E7-312F-4A15-8DD2-471C89CCB2D4 | ||
| Year | 2020 | ||
| Type of Item | Diploma Work | ||
| License |
|
||
| Bibliographic Citation | Angelos Kartakis, "Predicting the marketability potential of Google Play apps", Diploma Work, School of Electrical and Computer Engineering, Technical University of Crete, Chania, Greece, 2020 https://doi.org/10.26233/heallink.tuc.86704 | ||
| Appears in Collections |
In recent years there has been a great increase of interest around the field of smart phone application development growing to a $33B market. Applications that meet the various needs of customers are constantly evolving, leading to the development of online application marketplaces, such as Google Play. In this thesis, we examine the key drivers of app user ratings and propose four approaches to predict the ability of applications to be valued highly by their users within the app market, i.e. marketability. To develop these approaches, we leverage the predictive capacity of Machine Learning algorithms by formalizing the marketability prediction problem as a classification problem. In particular, we test and compare six Machine Learning algorithms --- i.e. Random Forests, Decision Trees, Multi-layer Perceptrons, k-Nearest Neighbor, Logistic Regression and Support Vector Machines --- for their ability to predict the ratings of app users based on a set of app features. By evaluating the algorithms against real data from Google play, we achieved up to 86% accuracy on marketability (i.e. user rating) prediction. The proposed solution can be extended to cover other domains, such as commercial capacity forecasting.
Copyright © DIAS 2013
