arXiv subjects

Emre Sefer

Publications and source records attributed to Emre Sefer.

3 recordsLinked to original sources

arXiv2026

Sub-City Real Estate Price Index Forecasting at Weekly Horizons Using Satellite Radar and News Sentiment

Reliable real estate price indicators are typically published at city level and low frequency, limiting their use for neighborhood-scale monitoring and long-horizon planning. We study whether sub-city price indices can be forecasted at weekly frequency by combining physical development signals from satellite radar with market narratives from news text. Using over 350,000 transactions from Dubai Land Department (2015-2025), we construct weekly price indices for 19 sub-city regions and evaluate forecasts from 2 to 34 weeks ahead. Our framework fuses regional transaction history with Sentinel-1 SAR backscatter, news sentiment combining lexical tone and semantic embeddings, and macroeconomic context. Results are strongly horizon dependent: at horizons up to 10 weeks, price history alone matches multimodal configurations, but beyond 14 weeks sentiment and SAR become critical. At long horizons (26-34 weeks), the full multimodal model reduces mean absolute error from 4.48 to 2.93 (35% reduction), with gains statistically significant across regions. Nonparametric learners consistently outperform deep architectures in this data regime. These findings establish benchmarks for weekly sub-city index forecasting and demonstrate that remote sensing and news sentiment materially improve predictability at strategically relevant horizons.

cs.LG↗

arXiv2026

Diagnosing LLM Reranker Behavior Under Fixed Evidence Pools

Standard reranking evaluations study how a reranker orders candidates returned by an upstream retriever. This setup couples ranking behavior with retrieval quality, so differences in output cannot be attributed to the ranking policy alone. We introduce a controlled diagnostic for reranking that uses Multi-News clusters as fixed evidence pools. We limit each pool to eight documents and pass identical inputs to all rankers. Within this setup, BM25 and MMR serve as interpretable reference points for lexical matching and diversity optimization. Across 345 clusters, we find that redundancy patterns vary by model: one LLM implicitly diversifies at larger selection budgets, while another increases redundancy. In contrast, LLMs underperform on lexical coverage at small selection budgets. As a result, LLM rankings diverge substantially from both baselines rather than consistently approximating either strategy. By reducing retrieval variance through fixed pools, we interpret these differences more directly as differences in ranking policy. This diagnostic is model agnostic and can be applied to any ranker, including open source systems and proprietary APIs. Our code and processed data for both the Multi-News diagnostic and the complementary TREC-DL evaluation are publicly available at https://github.com/barisarat/llm_reranker_multinews.git.

cs.LG↗

arXiv2021

BioCode: A Data-Driven Procedure to Learn the Growth of Biological Networks

Probabilistic biological network growth models have been utilized for many tasks including but not limited to capturing mechanism and dynamics of biological growth activities, null model representation, capturing anomalies, etc. Well-known examples of these probabilistic models are Kronecker model, preferential attachment model, and duplication-based model. However, we should frequently keep developing new models to better fit and explain the observed network features while new networks are being observed. Additionally, it is difficult to develop a growth model each time we study a new network. In this paper, we propose BioCode, a framework to automatically discover novel biological growth models matching user-specified graph attributes in directed and undirected biological graphs. BioCode designs a basic set of instructions which are common enough to model a number of well-known biological graph growth models. We combine such instruction-wise representation with a genetic algorithm based optimization procedure to encode models for various biological networks. We mainly evaluate the performance of BioCode in discovering models for biological collaboration networks, gene regulatory networks, metabolic networks, and protein interaction networks which features such as assortativity, clustering coefficient, degree distribution closely match with the true ones in the corresponding real biological networks. As shown by the tests on the simulated graphs, the variance of the distributions of biological networks generated by BioCode is similar to the known models' variance for these biological network types.

q-bio.MN↗