β campaign index Β· telemetry
| # | Candidate | FOSS implementation @ license | Paper Β· provenance | Coverage |
|---|---|---|---|---|
| #72 | psd_wiener_spectral_flatnessTonality-vs-noise descriptor of a bar's signal: ratio of the geometric mean to the arithmetic mean of the power spectrum of the mean-removed intra-bar π CRYPTO-PRIOR: present in patterns/crypto lit-scan (provenance only; still forex-proposable |
librosa/librosa (librosa.feature.spectral_flatness) ISC |
arXiv:1905.02284 git-anchor: absent @c803ef96e + @2d308f1c |
crypto-prior |
| #73 | psd_shannon_spectral_entropy_normNormalized Shannon spectral entropy: treat the normalized PSD of the bar's intra-bar tick path (or 200-bar close series) as a probability distribution π FOREX-COVERED: evaluated on forex_bars in the 2026-05-26/27 Gate campaign |
raphaelvallat/antropy (antropy.spectral_entropy, method='fft', normali BSD-3-Clause |
DOI:10.1016/0013-4694(91)90138 git-anchor: absent @c803ef96e + @2d308f1c |
β forex-dup |
| #74 | psd_spectral_centroid_normfreqSpectral center-of-mass ('brightness') of the bar's signal: the power-weighted mean frequency of the intra-bar mid tick path (or 200-bar close series) β forex-roadmap-mentioned (listed as an UNTESTED inventory item in the May Gate extended-inventory; not yet evaluated β ve π FOREX-MENTIONED-ONLY (roadmap inventory, not evaluated) |
librosa/librosa (librosa.feature.spectral_centroid) ISC |
HAL:hal-00692400 git-anchor: absent @c803ef96e + @2d308f1c |
not-covered |
| #75 | psd_log_spectral_slope_fullbandFull-band spectral slope: slope of an OLS line fit to log-power vs log-frequency across ALL positive frequency bins (DC excluded) of the bar's intra-b |
librosa/librosa (poly_features, order=1, over full band) ISC |
HAL:hal-00692400 git-anchor: absent @c803ef96e + @2d308f1c |
not-covered |
| #76 | midpath_zero_crossing_rate_meanremovedZero-crossing rate of the mean-removed intra-bar mid-price tick path: fraction of consecutive tick pairs where the de-meaned mid changes sign. A time- |
librosa/librosa (librosa.feature.zero_crossing_rate, threshold=0 ISC |
DOI:10.1109/PROC.1986.13663 git-anchor: absent @c803ef96e + @2d308f1c |
not-covered |
| #77 | schweizer_wolff_sigma_bidpath_askpathSchweizer-Wolff sigma: the normalized L1 distance between the empirical copula of (intra-bar bid-tick rank, ask-tick rank) and the independence copula |
cran.r-project.org/package=copBasic (wolfCOP) GPL-2.0 |
DOI:10.1214/aos/1176345528 git-anchor: absent @c803ef96e + @2d308f1c |
not-covered |
| #78 | hoeffding_phi_squared_midreturn_durationHoeffding's Phi-squared: the normalized L2 distance between the empirical copula of (mid log-return rank, bar-duration rank) and the independence copu |
cran.r-project.org/package=copBasic (hoefCOP) GPL-2.0 |
DOI:10.1016/j.jmva.2010.06.011 git-anchor: absent @c803ef96e + @2d308f1c |
not-covered |
| #79 | kendall_taub_signedreturn_spreadKendall's tau-b: the tie-corrected concordant-minus-discordant rank concordance between signed mid log-return and spread (bps) across the bar substrat |
scipy/scipy (scipy.stats.kendalltau, variant='b') BSD-3-Clause |
DOI:10.1093/biomet/33.3.239 git-anchor: absent @c803ef96e + @2d308f1c |
not-covered |
| #80 | gini_correlation_asymmetry_signedreturn_spreadSchechtman-Yitzhaki Gini correlation ASYMMETRY: the signed difference between the two directional Gini correlations R(return,rank(spread)) and R(sprea |
cran/GiniDistance GPL-2.0 |
DOI:10.1080/03610928708829359 git-anchor: absent @c803ef96e + @2d308f1c |
not-covered |
| #81 | energy_distance_bidpath_askpath_distributionSzekely-Rizzo energy distance between the empirical distribution of intra-bar bid tick increments and the empirical distribution of ask tick increment |
vnmabus/dcor MIT |
DOI:10.1016/j.jspi.2013.03.018 git-anchor: absent @c803ef96e + @2d308f1c |
not-covered |
| #82 | sign_negentropy_logcosh_gapHyvarinen robust negentropy of the standardized log-return series: the squared gap between the expected logcosh contrast on the data and on a matched |
scikit-learn (sklearn.decomposition.FastICA, _gsBSD-3-Clause |
DOI:10.1016/S0893-6080(00)0002 git-anchor: absent @c803ef96e + @2d308f1c |
not-covered |
| #83 | sign_two_state_entropy_rateShannon entropy rate of the return-sign sequence modeled as a two-state (up/down) Markov chain: estimate the up->down and down->up flip probabilities β lag-1 Markov sign-transition entropy (structural) Β· forex-roadmap-mentioned (listed as an UNTESTED inventory item in the π FOREX-MENTIONED-ONLY (roadmap inventory, not evaluated) |
github.com/gregversteeg/NPEET (entropy estimatorMIT |
DOI:10.3390/e23081046 git-anchor: absent @c803ef96e + @2d308f1c |
not-covered |
| #84 | empirical_cdf_fisher_rao_arclengthFisher-Rao (square-root-density / SRVF) arc length of the path traced by the empirical distribution of returns when read as a curve on the unit sphere |
github.com/jdtuck/fdasrsf_python (fdasrsf)BSD-3-Clause |
arXiv:1103.3817 git-anchor: absent @c803ef96e + @2d308f1c |
not-covered |
| #85 | updown_magnitude_cumulative_js_divergenceCumulative Jensen-Shannon divergence (Nguyen-Vreeken CJS) between the empirical distribution of |return| on up-bars and the empirical distribution of |
github.com/eda-rg/cjs (CISPA eda.rg 'cjs' refereMIT |
DOI:10.1007/978-3-319-23525-7_ git-anchor: absent @c803ef96e + @2d308f1c |
not-covered |
| #86 | interquote_count_fano_factorDispersion of the per-second (or per-bar-subinterval) quote counts relative to a Poisson process: variance-of-counts divided by mean-of-counts over th |
elephant-org/elephant (elephant.statistics.fanofactor) BSD-3-Clause |
DOI:10.3389/fncom.2020.569049 git-anchor: absent @c803ef96e + @2d308f1c |
not-covered |
| #87 | realized_semivariance_signed_ratioWithin-bar decomposition of realized variance into the part from up-ticks vs down-ticks of the mid: RS+ = sum of squared positive mid log-returns, RS- |
bashtage/arch Β· cran.r-project.org/package=highfrequency (realized measures); (rSVar) NCSA (arch) / GPL-2.0-or-later (hi |
DOI:10.1162/REST_a_00503 git-anchor: absent @c803ef96e + @2d308f1c |
not-covered |
| #88 | tickrule_run_length_entropyShannon entropy of the run-length distribution of consecutive same-direction tick-rule signs within the bar. Captures whether order flow comes in long β forex-roadmap-mentioned (listed as an UNTESTED inventory item in the May Gate extended-inventory; not yet evaluated β ve π FOREX-MENTIONED-ONLY (roadmap inventory, not evaluated) |
raphaelvallat/antropy (entropy primitives); RLE via numpy BSD-3-Clause |
arXiv:2003.05500 git-anchor: absent @c803ef96e + @2d308f1c |
not-covered |
| #89 | holden_effective_tick_clusteringEffective-tick (Holden/Goyenko-Holden-Trzcinka) clustering measure: from how often intra-bar prices land on coarse vs fine multiples of the symbol's m β needs symbol tick size (data constant); GPL bidask ref |
eguidotti/bidask (bidask: EDGE + companion low-freq spread proxie GPL-3.0 |
DOI:10.1016/j.jfineco.2008.06. git-anchor: absent @c803ef96e + @2d308f1c |
not-covered |
| #90 | intrabar_path_straightness_indexThe straightness (net-to-gross displacement) index of the intrabar mid-price path: net displacement (close - open in price, or Euclidean endpoint dist π CRYPTO-PRIOR: present in patterns/crypto lit-scan (provenance only; still forex-proposable |
JimMcL/trajr (TrajStraightness) GPL-3.0 |
DOI:10.1016/j.jtbi.2004.03.016 git-anchor: absent @c803ef96e + @2d308f1c |
crypto-prior |
| #91 | intrabar_path_total_absolute_turningSum of the absolute exterior turning angles at every interior vertex of the intrabar (time,price) tick polyline, normalized by total arc length (total |
JimMcL/trajr Β· scipy/scipy (TrajAngles / turning angles); (vector algebra) GPL-3.0 |
DOI:10.1145/1090122.1090160 git-anchor: absent @c803ef96e + @2d308f1c |
not-covered |
| #92 | intrabar_gyration_tensor_asphericityForm the 2x2 radius-of-gyration tensor of the intrabar point cloud {(t_i, p_i)} (tick mids, mean-centered, in normalized time-price coordinates), take |
glotzerlab/freud Β· plumed/plumed2 (ClusterProperties gyration tensor); (GYRATION BSD-3-Clause |
DOI:10.1063/1.1674820 git-anchor: absent @c803ef96e + @2d308f1c |
not-covered |
| #93 | intrabar_convex_hull_area_to_pathlength_ratioCompute the convex hull of the intrabar (time,price) tick point cloud in normalized coordinates; report the dimensionless ratio hull_area / (path_arc_ |
scipy/scipy (scipy.spatial.ConvexHull / Qhull) BSD-3-Clause |
DOI:10.3390/ijgi7010014 git-anchor: absent @c803ef96e + @2d308f1c |
not-covered |
| #94 | intrabar_normalized_levy_areaThe LΓ©vy area of the intrabar 2D path lift (e.g. (time, mid) or the (bid, ask) co-path), i.e. the signed area enclosed between the path and the chord β level-2 path-signature fixed depth (structural) |
bottler/iisignature Β· datasig-ac-uk/RoughPy GPL-3.0 |
arXiv:1603.03788 git-anchor: absent @c803ef96e + @2d308f1c |
not-covered |
| #95 | intrabar_h0_total_persistenceSublevel-set persistent homology (H0) of the intrabar mid-price function vs. time: sweep the price level upward, record birth (local minima) and death π CRYPTO-PRIOR: present in patterns/crypto lit-scan (provenance only; still forex-proposable |
scikit-tda/ripser.py Β· giotto-ai/giotto-tda (CubicalPersistence/lower-star) MIT |
arXiv:2311.01115 git-anchor: absent @c803ef96e + @2d308f1c |
crypto-prior |
| Candidate | Drop reason |
|---|---|
intrabar_divider_fractal_dimension | gate: not-parameterless |
refined_unbiased_distance_correlation_signedreturn_spread | gate: forex-dup:bid_ask_intrabar_distance_correlat |
quadrant_count_ratio_mean_split_signedreturn_spread | gate: not-parameterless |
cepstral_peak_prominence | gate: not-parameterless |
micro_price_vs_mid_divergence | gate: not-computable(PARTIAL) |
roll_serial_covariance_spread | gate: forex-dup:edge_spread_bps (edge is an effici |
goh_barabasi_burstiness_coefficient | gate: forex-dup:cv2_consecutive_intervals (b=(cv-1 |
quote_lifetime_gini | gate: forex-dup:gini_interquote_interval (same obj |
symbolic_transfer_asymmetry_lag1_signs | gate: not-parameterless |
fisher_information_nongaussianity_gap | gate: not-parameterless |
kendall_s_signs_serial_concordance | gate: forex-dup:mann_kendall_s_z (mann-kendall is |
psd_spectral_spread_p2 | internal-dedup: 2nd spectral moment, partner of centroid (kept) β high redundancy |
psd_spectral_crest_factor | internal-dedup: peak/rms β inverse of spectral flatness (kept) |
psd_dominant_frequency_power_ratio | internal-dedup: dominant-freq concentration β flatness/crest |
spearman_footrule_signedreturn_spread | dup-class: footrule already dropped iter-2; dominated by Kendall/Spearman |
hbkr_b_statistic_empcopula_bidpath_askpath | internal-dedup: L2 empirical-copula independence β hoeffding_phi_squared (kept) |
gini_gamma_cograduation_bidpath_askpath | internal-dedup: monotone rank β kendall (kept); keep gini_correlation_asymmetry instead |
sign_run_record_growth_moment | near-dup vs iter-1 foster_stuart_records_d (record statistics) |
sign_string_bdm_complexity | near-dup vs iter-1 lziv_complexity_signs + borderline (BDM block size/CTM base) |
updown_magnitude_energy_distance | internal-dedup: same up-vs-down magnitude divergence as updown_magnitude_cumulative_js_divergence (kept) |
duration_index_of_dispersion | internal-dedup: var/mean of durations β interquote_count_fano_factor (kept) |
intrabar_path_isoperimetric_quotient | internal-dedup: area-compactness β intrabar_convex_hull_area_to_pathlength_ratio (kept) |
intrabar_birnbaum_saunders_shape_alpha | cluster-mismatch: parametric distribution-shape fit, overlaps iter-1 L-moment/quantile shape family |
tickrule_price_impact_slope | forex-prior: Kyle-lambda analog β kyle_lambda_proxy_q1000 IS a forex_bars column (data_types.rs:398) |
realized_amihud_per_quote | forex-prior: Amihud illiquidity already in pipeline + evaluated (amihud_illiquidity / amihud_illiquidity_q1000, data_types.rs:264/369; iter#84-86 chronicles) |