Mixed Causal Structure Discovery with Application to Prescriptive Pricing

Authors: Wei Wenjuan, Feng Lu, Liu Chunchen

IJCAI 2018 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Experiments on simulate/real datasets show the superiority of our new approach in both price-demand law recovery and demand forecasting, as well as show promising performance to support optimal pricing.
Researcher Affiliation Industry NEC Labs China wei wenjuan@nec.cn, feng lu@nec.cn, liu chunchen@nec.cn
Pseudocode Yes Algorithm 1 A* Fo Ba with ancestor-based space cutting
Open Source Code No The paper does not provide any links or explicit statements about making the source code for the described methodology available.
Open Datasets No We simulated prices and sales of 40 kinds of beer, and 9 objective factors (i.e., temperature (continuous), seasonality (4, binary), vacation (binary), promotion (binary) and sunny/rainy (2, binary)). We applied our method to real retail data from a supermarket in Tokyo. The data is provided by KSP-SP Co.,LTD.
Dataset Splits Yes We simulated 16 experiment settings differed in noise types, sparsity (simple/complex structure), train sample scale (1/3 years), and test sample scale (7/30 days). To evaluate the pricing strategy, we built an independent validation environment ˆl ( ) by estimating new sale formulas on data from 2013/01 to 2013/09. Compared to ˆl( ), ˆl ( ) is an environment that is more similar to the real one in the validation stage (2013/10).
Hardware Specification No The paper does not provide specific hardware details (exact GPU/CPU models, processor types, or memory amounts) used for running its experiments.
Software Dependencies No The paper mentions various algorithms and methods (e.g., 'Fo Ba algorithm', 'l0 sparse regression', 'stable PC') but does not provide specific version numbers for any software dependencies or libraries.
Experiment Setup No The paper mentions using 'default parameter settings' for benchmarks and describes how synthetic data coefficients were generated, but it does not specify concrete hyperparameter values or detailed training configurations for its own proposed method.