SuperAPI AI Directory
Partner with us中文AI newsFavorites
Skillslight-venue-matching

light-venue-matching

Data & analytics

View repository

Updated 10-16 06:23

About this skill

Submission positioning and venue matching

Assessment input (first take stock of strengths)

Paper quality (innovation/experimental completeness/theoretical depth), research direction, language level, author background (undergraduate/master's/doctoral, whether there is a strong advisor), time requirements (deadline/graduation requirements), budget (whether APC can be paid), goals (graduation/awards/job hunting/pure publication).

Matching process

  1. Use the db01 journal and conference database to filter candidates by direction (subject_area); when insufficient, programmatically expand candidates by direction: OpenAlex Sources GET https://api.openalex.org/sources?filter=topics.id:<方向>,is_oa:<bool>&sort=cited_by_count:desc&per_page=50&mailto=<email>(cursor cursor=* pagination, can filter by summary_stats.2yr_mean_citedness/apc_usd/is_in_doaj ); for computer science directions, directly take candidates from the CCF catalog by "field + target tier (A/B/C)".
  2. Align venue level with paper strength—do not only recommend "prestigious" ones. For students, use "tiers" (CCF A/B/C, CAS quartiles, Peking University Core/CSSCI) rather than raw IF.
  3. For each candidate fill inunified comparison fields(sources to retrieve are shown in parentheses):
    • Impact metrics: IF + JCR quartile (JCR/LetPub), CiteScore + percentile (Scopus), SJR/Eigenfactor reputation (Scimago/IEEE), CAS quartile major category/minor category (quartile table/LetPub).
    • Tier: CCF A/B/C, Peking University Core/CSSCI/CSCD inclusion status.
    • Cycle: time to first decision + weeks from submission to publication (JournalFinder/IEEE Recommender/LetPub).
    • Fees and OA: APC, whether OA, whether indexed in DOAJ (DOAJ API GET /api/v3/search/journals/issn:<ISSN>)。
    • Acceptance likelihood and background:Officially public acceptance rate(fill in a specific number and attach a link only when disclosed on the journal/conference official website or in a formal report; otherwise fill in "to be verified—no official public data"), proportion of Chinese authors (LetPub, marked as a community experience estimate), self-citation rate, whether it suits the author's background, template and citation format (reference_style).Do not retrieve or fill in LetPub's "acceptance ratio" as a probability number(it self-acknowledges being a community submission experience estimate, not official statistics); use only the rubric below for qualitative grading.
    • Indexing verification: whether indexed by WoS and the index (SCIE/SSCI/AHCI/ESCI, check Master Journal List), whether indexed by Scopus. The three quartile systems (JCR/SJR/CAS) have different criteria; each item must be labeled with source + year and must not be mixed.
  4. Early warning screening (bidirectional whitelist + blacklist)：
    • Positive whitelist signals: indexed in DOAJ (especially with DOAJ Seal), formally indexed by WoS/Scopus.
    • Blacklist/predatory characteristics: matches the CAS International Journal Early Warning List (high/medium risk), or abnormal self-citation, ultra-fast review, surge in annual publications, abnormally high proportion of Chinese authors, high APC, fake metrics. If matched, mark red and advise against (linked to a10).

Tiered recommendations (must give three tiers)

  • Reach: a stretch but possible, high return.
  • Target: matches strength, high probability of acceptance.
  • Safety: ensures publication/graduation. Give 1–3 for each tier, with reasons,Qualitative acceptance likelihood grading (High/Medium/Low, see rubric below; do not fabricate percentages), timeline, cost, risk.

Acceptance likelihood assessment rubric (qualitative grading; fabricating probability numbers is prohibited)

Iron rule: Unless the journal/conference's official website or a formal report publicly discloses an acceptance rate and a link can be attached, otherwisenever give an exact acceptance probability/percentage. The "acceptance ratio" on aggregator sites such as LetPub is a community submission-experience estimate, not official statistics,and must not be cited as a probability number. In the end, output only a qualitative grading of "High/Medium/Low + item-by-item reasons".

Five verifiable signals (score each item, each as High/Medium/Low, and mark the source)

  1. Author's relative strength: the author's h-index over the past 5 years / citations of representative works (OpenAlex Authors GET https://api.openalex.org/authors?search=<姓名>&mailto=<email>, take summary_stats.h_index) compared with the venue's summary_stats.h_index(already stored in db01 or looked up via OpenAlex Sources). If the author's influence is close to/exceeds that of typical authors in the journal → High; clearly lower → Low. Note: for h-index, same-name authors need to be disambiguated using institution/ORCID; mark as "to be verified" rather than forcibly asserting.
  2. Direction match: the degree of topic overlap between the paper's subject and the venue's subject_area(db01) and recent representative_papers(db01 field). Core topic hit → High; borderline/interdisciplinary → Medium; clearly deviating from the journal's scope → Low.
  3. Method/data scale match: whether the methods, dataset scale, and experimental volume used in the paper meet the threshold reflected by the venue's representative_papers (e.g., top conferences often require large-scale experiments + SOTA comparisons; field journals look at whether there is a method paradigm preferred by that journal). Meets the standard → High; partially meets it → Medium; clearly insufficient → Low.
  4. Official acceptance rate tier: only when an official public acceptance rate is available, assign the tier by value (e.g., <15% recorded as extremely competitive, 15–30% high, >30% medium);if there is no official data, fill this item as "to be verified—no official acceptance rate", do not participate in fabricating numbers, and explain in the conclusion that this item is missing.
  5. Self-assessed innovation: have the author/assessor self-assess according to the three tiers "incremental improvement / significant improvement / new problem or new paradigm", against the venue's tier (top conferences/top journals prefer the latter two tiers). This is a subjective item and must be explicitly marked "author self-assessment, not an objective indicator".

Aggregation rules

  • If most of the five items are "High" and there is no fatal weakness (failure to meet the standard in direction/method counts as fatal) → overall rating High (safe/fallback candidate)。
  • If the signals are mixed high and low, with 1 fatal weakness → overall rating Medium (reach candidate; explain the weakness)。
  • If the direction or method clearly fails to meet the standard, or the author's strength is far below that of the journal → overall rating Low (not recommended, or only as an alternative outside the fallback options)。
  • For an output format example, see item 1 of "Output" below; each grading must be followed by "because… (which signal is cited + source)", and must not give only the conclusion.

Tool/data perspective (each performs its own role; do not mix standards)

  • Programmatically pull candidates/metadata: OpenAlex Sources (free REST,api.openalex.org/sources, filter by ISSN/topic/OA/APC/h-index, cursor pagination, add mailto to the polite pool).
  • Impact metrics: JCR (JIF, 2-year window, currently 1 decimal place, unified SCIE+SSCI+ESCI ranking since 2024, subscription required) | Scopus CiteScore (4-year window, denominator includes conference papers/reviews, free) | Scimago SJR (PageRank-like prestige weighting, Scopus data, Excel downloadable from portal) | Eigenfactor/Article Influence (built into IEEE Recommender).
  • Indexing verification: WoS Master Journal List (free check of whether indexed + SCIE/SSCI/AHCI/ESCI) | Scopus Source List | EI verification: authoritative source = Engineering Village official Compendex Source List(retrieve the current Excel from the product page "View source list", search by journal name/ISSN; strictly prohibited to cite outdated third-party "EI source journal" sites), see references.md "EI Verification Path".
  • Legitimacy of domestic conferences: CCF catalog (computer science) tiering + verify organizer on first-level society official website + recheck publication/indexing destination; three red lines for fake conferences (pay-to-guarantee-acceptance/indexing promise with no source/impersonated organizer), see references.md "Domestic Conference Signal Sources".
  • Domestic tiers: CCF A/B/C catalog (top ten computer science fields) | CAS partition (major/minor categories, Zone 1 approx. top 5%, with early warning list) | PKU Core (Chinese core, new edition about every 4 years) | CSSCI (NJU Core, social sciences) | CSCD (science and engineering).
  • Manuscript-journal matching tools(only discover candidates, with publisher bias): Elsevier JournalFinder, Springer Journal Suggester, IEEE Publication Recommender—enter title/abstract, returns journals under the same publisher + cycle/acceptance rate/APC/OA.
  • One-stop China-friendly profile: LetPub (IF/JCR quartile/CAS partition/review cycle/acceptance rate/proportion of Chinese authors/APC/early warning all in one place, but key values must be re-verified against official sources).
  • OA compliance: DOAJ (free API doaj.org/api/v3/search/journals/{query}, inclusion = positive signal, having the Seal is even better).

Metadata is based on db01 and marked with last_checked_date; specific endpoints/parameters/pitfalls see references.md.

Script: five-signal cross-check query (scripts/venue_signal.py）

When evaluating a candidate venue, first run this script to get the five-signal cross-check JSON, then use it to fill the verifiable signal columns of the comparison table—faster than manual checking and with unified criteria:

# 期刊 + 作者：五信号全开（作者匹配度需 --author）
python scripts/venue_signal.py --issn 1234-5678 --author "Zhang San" --venues-csv <db01路径>/venues.csv
# 仅期刊：信号1/2/5（审稿周期/分区从 db01 卡取）
python scripts/venue_signal.py --issn 1234-5678 --card-fields review_cycle="约8周" apc_fee="1800 USD" cas_quartile="2区"
  • Output five signals: ① publication volume trend (counts_by_year) ② rough self-citation rate check (outward self-citation, proportion of citations to this journal by this journal) ③ review cycle (db01 card) ④ overlap between authors and the journal's topics ⑤ APC and partition (db01 card + OpenAlex).
  • Each signal degrades independently: if data retrieval fails, mark status:"unavailable"+reason, do not fabricate numbers; output as much as can be checked.
  • Author match degreetake the first hit by name, without disambiguation, output with disambiguation_caveat, must be secondarily confirmed via ORCID/institution, corresponding to the h-index same-name disambiguation discipline of the rubric "relative author strength".
  • Self-citation rate is anoutward self-citation rough estimate(not the official incoming journal self-citation rate); the output has noted the method; if it hits >25%, prompt for early-warning screening comparison (abnormal self-citation).
  • OpenAlex access specifications (key/rate limits/billing) see the "OpenAlex access truth source" section in m01 references; the script does not duplicate the numbers.

Output

  1. Candidate venue comparison table (including the above fields + recommended tier + qualitative grading of acceptance likelihood). Table skeleton see templates/venue_compare_table.md, filling the table ensures no field is omitted and criteria are unified. Example of writing the grading column: 中｜方向匹配=高(主题命中CVPR scope,db01 subject_area=计算机视觉);作者实力=中(作者h-index=18 vs 该会代表作者多>40,OpenAlex Authors);方法规模=中(单数据集,顶会偏好多数据集SOTA);官方接收率=待核查(CVPR官网未稳定公开逐年录用率);创新性=作者自评"显著改进"(主观) —— Note:No percentages throughout, "to be verified" is retained as is.
  2. Submission strategy recommendations (which to submit to first, order of resubmission after rejection).
  3. Risk alerts (warnings/long cycle/high fees/anonymity requirements).

Connection

After selection → m12 applies the corresponding template, m10 adjusts the citation format; submission records and decisions go into db09; when resubmitting after rejection, return to this skill to reorder. All journal data must be re-verified before submission (CONVENTIONS §1).

For real tool endpoints/parameters/known pitfalls, see references.md in the same directory.