Online Search with Best-Price and Query-Based Predictions
Authors: Spyros Angelopoulos, Shahin Kamali, Dehou Zhang9652-9660
AAAI 2022 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We also provide experimental results on data obtained from stock exchange markets that confirm the theoretical analysis, and explain how our techniques can be applicable to other learning-augmented applications. |
| Researcher Affiliation | Academia | Spyros Angelopoulos1, Shahin Kamali2, Dehou Zhang2 1 Centre National de la Recherche Scientifique (CNRS) 2 University of Manitoba, Winnipeg, Manitoba, Canada |
| Pseudocode | No | The paper describes algorithms but does not provide structured pseudocode or algorithm blocks. |
| Open Source Code | Yes | 1https://github.com/Dehou Zhang/Online-Search-with Predictions and https://github.com/shahink84/Online Search RBIS |
| Open Datasets | Yes | We evaluate our algorithms on benchmarks generated from real-world currency exchange rates, which are publicly available on several platforms. Specifically, we rely on (EA Trading Academy 2021). |
| Dataset Splits | No | The paper mentions generating instances of the problem but does not specify training, validation, or test dataset splits. |
| Hardware Specification | No | The paper does not explicitly describe the hardware used for running its experiments. |
| Software Dependencies | No | The paper does not provide specific software dependencies with version numbers. |
| Experiment Setup | Yes | We select 500 values of negative error equally distanced in [0, 0.5], as well as 500 equally-distanced values of positive error in [0, 0.5]. For each selected value, say η0, and for each instance Ix of the problem, we test our algorithms for prediction error equal to η0, that is, the predicted value is generated by applying the error η0 on the best price in Ix. We evaluate ORAr with different values of the parameter r {0.5, 0.75, 1.0, 1.25, 1.5}. We tested ROBUST-MIX with upper bound H on both the positive and negative error, that is, H = Hn = Hp for H {0.1, 0.2, 0.3, 0.4, 0.5}. In our experiments, we set the number of queries to n = 25. We test RLIS and RBIS with H taken from {3, 5, 8, 10, 13}. |