**4. Conclusions and future work**

The huge amounts of data, faced when analyzing transiting exoplanet light curves, have encouraged data scientists to develop machine learning models capable of automatically identifying exoplanets. These models can reduce the time spent

eyeballing the light curves while enhancing the identification accuracy. For such algorithms to exist, simulated light curves are necessary, because they provide a wide variety of labeled scenarios that can be used to train the models. For this reason, in this work, we presented the methodology followed to create two datasets of simulated light curves with different parameters, labeled as transit and nontransit signals. These light curves were used to train machine learning algorithms, and later test them. Once that the results obtained with the simulated data are satisfying enough, real data can be used to identify transiting exoplanets and contribute to the existing catalogs of exoplanet discoveries. Furthermore, some useful preprocessing steps were explained in this work. They can be used with simulated or real data. Our results show that using the multiresolution analysis techniques to preprocess the light curves improves the identification rates of the machine learning models. Future work will be done in proposing a new machine learning model based on multiresolution analysis techniques, instead of using them to preprocess the light curves.
